Skip Navigation.
Section 0

Hardware Fault Diagnosis & Repair

by Martin Straw [21010093+]

Version 2.34

Copyright: 1996-1997 Martin Straw. All rights reserved.

Warning: Please remember that when the lid of the Chroma is lifted, the power supply board is easily accessible, both deliberately and accidentally - please be careful and also ensure that the instrument and your tools are adequately earthed. The Chroma is a highly technical device, which should only be repaired by fully qualified personnel.

Disclaimer: Although every effort has been made to ensure that the information in this document and related images is correct, the author cannot be held responsible for any inconvenience, loss, damage or harm, arising from following the advice or data held herein. It is the reader's responsibility alone to decide whether and how to use this document and related images and to avoid any problems, whether physical, financial or of any other kind.

Contents

(1) Introduction

This document explains how to diagnose and repair faults in the ARP Chroma and Rhodes Chroma synthesizers, but NOT the Chroma Polaris.

The information has been collected and deduced over a number of years, in the course of successful repairs of a number of Chromas.

Note that the Chroma will often break if switched on at a temperature that is any less than eighteen degrees Celsius. Sixteen degrees is dodgy and fourteen degrees centigrade often causes some kind of failure.

Temperature problems are usually of two types: a CMOS or TTL logic chip fails, often creating a dead short, or less frequently, a capacitor fails, also with a dead short. Elderly TTL chips seem to be more robust than elderly CMOS.

There are two main types of Chroma hardware faults:

(2) Faults That Affect All Of The Chroma's Voices

(2.1) Eight Most Common Faults

(2.1.1) Progressive tuning failure

The instrument works for a very short time, but when AUTO TUNE is pressed, several voices fail, producing an error message such as "Err 0457" in the DATA READOUT display.

Pressing AUTO TUNE again results in more voices failing, until they all fail, producing the error message "E01234567".

This is almost certainly the power supply board failing as it warms up, causing the plus and minus 12V rails to become unequal. They should be equal, but of opposite sign.

Note that although severe cases will cause "E01234567" to be displayed as soon as the Chroma is switched on, the immediate failure of all voices can also be caused by several other failures, such as:

Details of diagnosis and repair are given later in this document.

(2.1.2) The Chroma responds to buttons but ignores the keyboard

Pressing buttons on the control panel changes the DATA READOUT and LEDs correctly, but playing the keyboard produces no sound. If a MIDI interface is fitted, MIDI input also produces no sound.

There are three likely causes of this fault:

Details of diagnosis and repair are given later in this document.

(2.1.3) The Chroma appears to be switched on but ignores all input

After working correctly for a few minutes, the Chroma still appears to be switched on, but ignores button and key presses. If a MIDI interface is fitted, it also ignores MIDI input.

If this fault is allowed to continue without repair, the time between power-on and failure reduces to almost zero, resulting in symptoms described in section (2.1.4) or (2.1.5).

The power supply board is probably failing, with the plus 5V rail sagging a little below 5 Volts, causing the computer board to fail first, since this is always operating at a slightly reduced voltage, due to its distance from the power connection.

Details of diagnosis and repair are given later in this document.

(2.1.4) The Chroma has all LEDs illuminated but ignores all input

When switching on the Chroma, the LEDs begin to flash, but soon freeze with them all illuminated, or it immediately freezes with all LEDs illuminated. The synthesizer does not respond to button or key presses. If a MIDI interface is fitted, MIDI input is ignored. This is a more severe version of problem (2.1.3).

The power supply board is probably failing, with the plus 5V rail sagging a little below 5 Volts, causing the computer board to fail first, since this is always operating at a slightly reduced voltage, due to its distance from the power connection.

However, this problem can also be caused by a failure in the keyboard scanning circuit on the stack switch board, or dirty connections from between the I/O board and the computer board.

Details of diagnosis and repair are given later in this document.

(2.1.5) The Chroma appears to switch itself off

The Chroma works correctly for a few minutes, but when it has warmed up, the LEDs and displays suddenly go out, as though the power has been switched off. Switching the instrument off and on makes it work for a few seconds.

This is a more severe version of problems (2.1.3) and (2.1.4).

The power supply board is probably failing, with the plus 5V rail sagging a little below 5 Volts, causing the computer board to fail first, since this is always operating at a slightly reduced voltage, due to its distance from the power connection.

Details of diagnosis and repair are given later in this document.

(2.1.6) Parameter values jitter

The Chroma appears to work correctly, except when adjusting a program parameter which has a high-definition range, such as the -64 to +63 range of the pitch modulation depth parameter buttons 28, 30 and 32.

When you let go of the PARAMETER CONTROL slider, the value shown in the DATA READOUT display jitters between two settings, such as 4 and 5.

You also notice, when listening very carefully, that the entire pitch of the instrument is also jittering up and down. This is an extremely slight pitch change, which some people cannot hear.

Once again, this is caused by the power supply board producing an unstable plus 5V output, causing the DAC and ADC circuits to jitter.

(2.1.7) Program parameters are forgotten or set to random values

If program parameters are forgotten, or set to random values, whilst the Chroma it is still switched on, it could be the power supply board or the CMOS RAM chips at fault.

If program parameters are forgotten, or set to random values, after the Chroma has been switched off, it could be the batteries or the power supply board reset circuit at fault.

See section 4 for full details.

(2.1.8) All voices produce defective sound

All voices produce sound, but it is defective, not as defined by the patch program parameters.

In mild cases, only very complex programs produce defective sound. In bad cases, all programs produce defective sound.

This could be the power supply board being unable to deliver a full 5V supply to the computer board, or an op-amp on the EQ or mother board.

If the power droops a little, to around 4.88V, the processor can often cope with all but the most complex voice programs. As the voltage droops further, the effect is more serious, until the Chroma fails completely, with symptoms described in sections (2.1.3) through (2.1.5).

(2.2) Strategy For Fault Diagnosis

If a fault affects all voices, it is almost certainly caused by deterioration of the power supply board, but a number of other faults can produce the same effects, so these should be checked and eliminated before moving on to the time-consuming task of repairing the power supply unit.

You should check the circuit boards in the following order:

  1. Power supply board
  2. Computer board
  3. EQ board
  4. Mother board
  5. Keyboard scanner and stack switch board
  6. I/O board
  7. The voice boards

This document has a separate section for each board, giving details of diagnosis and repair.

(3) CHECKING THE POWER SUPPLY BOARD

The power supply board should be checked first, because it operates at a much higher temperature than the rest of the instrument, causing it to fail most frequently, resulting in a wide variety of errors.

Check all the power rail output voltages with a digital voltmeter. The output wires are color coded as follows:

0V analogue BLACK not too far above 0V
0V digital GREY not too far above 0V
0V analogue WHITE ground for tapper solenoid
+5V battery BLUE drops to 0V to engage the CMOS memory batteries
+5V analogue YELLOW adjust to +5.05V, via pot R22, for best reliability
+5V digital ORANGE adjust to +5.05V, via pot R22, for best reliability
+12V analogue RED must be equal value, but opposite sign, to -12 rail
-12V analogue VIOLET must be equal value, but opposite sign, to +12 rail
+24V analogue BLUE for tapper solenoid, only needs to be between 20-30V

When checking the 12V outputs, note that a Chroma running at +12.06 and -12.06 Volts works perfectly, but a Chroma running at +12.06 and -11.56 Volts will not be able to tune any voices at all. They must be almost exactly equal but opposite.

When checking the 5V outputs, note that some MIDI interfaces take their power from the Chroma's COMPUTER INTERFACE socket, raising the temperature of the power supply board.

Try unplugging any device attached to the COMPUTER INTERFACE socket, and see if the synthesizer works normally for at least an hour, in which case the the external device is probably faulty, drawing too much power.

You must also perform an additional check of the digital +5V output, by measuring the voltage on the computer board.

The computer board, underneath the right hand side of the control panel, is furthest from a direct feed from the +5V digital power rail, obtaining its power via the huge I/O board, which is mounted under the center of the control panel. You should therefore measure the voltage on the computer board, as well as the PSU, to check the +5V digital power rail.

Even though you can adjust the +5V rail by pre-set potentiometer R22, you will find that by the time the power has reached the computer board, the voltage has sagged. If the computer board falls below 4.89V the Chroma will begin to fail, and the power going into capacitor C3 must be increased, as described in section 3.1.3.

A particularly good computer board may be able to partly function as low as 4.84V, but any less than that will probably result in complete failure.

If the power supply board appears to be failing, note that faulty voice boards often affect the power output. The CMOS chips are particularly likely to fail with an internal short, or low resistance, causing the power rails to drop below their proper value.

If one of the power rails shows a low voltage, take out the voice boards one at a time to see if one board causes the instrument to fail. If removing a particular board cures the fault, replace all the other boards to check that it is not simply the quantity of boards being powered that the supply unit cannot cope with.

(3.1) Repairing The Power Supply Board

See Tom Moravansky's [21030431] September 2006 ChromaTalk post Martin Straw's power supply update procedures for some additional instructions and observations.

Removing and repairing the power supply board is a time-consuming and labour intensive task. It is therefore very strongly recommended that all of the modifications listed below should be done at the same time, as one big task.

The changes will make the board last for many years without further repair and will also make the tuning so stable that, after the initial warm-up, it will behave like a digital device.

(3.1.1) Replacing nearly all transistors and op-amps

An un-modified Chroma power supply board generates too much heat, which slowly breaks down the transistors and op-amps over a few years.

Having taken the trouble to take the PSU out of the instrument, the first thing to do is replace the blanking panel, next to the COMPUTER INTERFACE, with a wire grill that allows the air to flow through it.

Next replace all transistors, except the big 2N6258, in the TO3 case, which I have never known to fail. This sounds like a big job, but there are only five power transistors in TO220 casings and seven general-purpose transistors in TO92 cases.

Pages 6-16 and 6-17 of the Rhodes Chroma [Service] manual clearly show all aspects of the PSU, but in case these are not available, here is a description of how to find your way around the board.

At the back of the PSU, the power transistors are screwed to the iron heat- sink, which also holds all the jack-sockets, the power switch and the COMPUTER INTERFACE D-connector.

Starting from the left, nearest the power supply switch, the power transistors to be replaced are:

The general-purpose transistors are all 2N3904 or 2N3906. Starting from the left, that is from the same end as the power supply switch, the transistors to be replaced are:

Next, replace the two LM358 op-amps Z1 and Z2. Z1 is found in front of Q6(TIP29A) and Z2 is some way in front of Q7(TIP110) next to the 5V adjust pre-set potentiometer R22 (10K).

Do NOT solder the op-amps into position. Solder wire-wrapping sockets in place instead, leaving them standing as far above the board as possible. This acts as a further heat-sink and also allows easy replacement without removing the board from the Chroma.

(3.1.2) Modifying the plus and minus 12V circuits

The +12V and -12V rails are the most frequent cause of tuning failures.

Resistors R26 and R31 are specified as 1 Watt, which is too low. Over a period of time, the heat causes them to break down internally, altering their resistance, causing the +12V and -12V rails to drift. These rails MUST be of equal value, but opposite sign.

A Chroma running at +12.06 and -12.06 Volts works perfectly, but a Chroma running at +12.06 and -11.56 Volts will not be able to tune any voices at all.

If the plus and minus 12 Volt rails are unequal, resistors R26 and R31 should be replaced with 4 Watt, 1 Ohm, wire wound resistors, of 10 percent or better tolerance. These are extremely reliable, but it is also advisable to replace the high-precision, 1 percent, 12.1K resistor R27 with two wires going to a 2 percent 11K Ohm resistor, in series with a 2K Ohm preset potentiometer. This will allow the minus 12 Volt rail to be finely adjusted to balance the plus 12 Volt rail.

R26 and R31 are easy to find, because they are fat resistors, on the edge of the board, next to the voice cards. The 12.1K R27 is between the pre-set potentiometer, that adjusts the 5 Volt rail, and R31.

R25 and R30 have also been under-specified. They last much longer than R26 and R31, but they will still fail after a few years and should be replaced with 2 Watt, 1K Ohm, 5 percent tolerance resistors.

R25 and R30 are also easy to find, because on an un-modified power supply board, they are the only 1 Watt resistors close to R26 and R31.

R25 is the only resistor in this area which is oriented at 90 degrees to all the other resistors, next to the 2N3904 transistors Q8 and Q10.

R30 almost touches both R26 and R31, being mounted just to the side of the gap between them.

A picture of the relevant section of the power board, with relevant notes, is given in the following diagram.

  1. The original R26 and R31 have such a low power rating that the heat causes them to break down internally, causing the plus and minus 12V rails to drift away from being equal but opposite. Replace them with 1 Ohm 4 Watt 10% wire wound resistors.
  2. Similarly, R25 and R30 should be replaced with 1K 2 Watt 5% resistors.
  3. R27 (12.1K) balances the plus and minus 12 volts rails and should be replaced with flying leads to an 11K 0.25 Watt 2% resistor in series with a 2K pre-set potentiometer to allow fine adjustment.
Q7 = TIP110 R23 = 4.99K R29 = 3.3K
8 = 2N3904 R24 = 6.98K R30 = see text (2)
Q9 = TIP110 R25 =see text (2) R31 = see text (1)
10 = 2N3904 R26 = see text (1) R32 = 10K
11 = 2N3906 R27 = see text (3) R33 = 470
13 = diode 1N34A R28 = 12.1K Z2 = LM358

I recommend putting the new 11K resistor and 2K potentiometer on a small new board, along with other components required to modify the pre-regulator circuit. The circuit diagram and notes for the new board are shown in Addition to Pre-Rectifier Circuit and Addition to +12V and -12V Regulator below.

I recommend attaching the new board by drilling into the back of the PSU heat-sink, next to the power switch, then screwing small mounting brackets into position, so that the new board is mounted parallel to the main PSU board.

(3.1.3) Modifying the pre-regulator circuit

Some power supply boards were modified at the Chroma factory in a way that invalidates connections F1, F2 and F3. This would appear to be to overcome deficiencies in the board's performance. Please check the board carefully before proceeding with the modification.

The first few Chromas had an adjustment potentiometer in the pre-regulator circuit, allowing adjustment of the power supplied to the huge vertically mounted capacitor C3.

C3 is a 7.5V, 47000uF capacitor, with a crow-bar over-voltage protector mounted across its terminals. The protector looks very much like a power transistor in a TO3 case. Capacitor C3 supplies power to the +5V digital linear regulator and therefore indirectly powers all the TTL logic chips and many other logic devices.

The pre-regulator circuit needs the adjustment potentiometer, because the level of power to C3 is crucial, but somewhere between serial numbers 21000001 and 21010093, the PSU board tracks were altered, so that soldering the potentiometer and related components into position would not only fail to work, but would actually damage the PSU, if it was switched on.

If you exceed 7.5V going into C3 by too much, the capacitor will fail, hence the crow-bar over-voltage protector to prevent accidental damage. In fact, some Chromas only need peaks of 6.8V fed into the terminals of C3, to store enough power to feed the TTL chips on the I/O and computer boards. At the other extreme, some Chromas need peaks of 7.4V across C3 to make them run smoothly.

A classic fault of the Chroma is that, after a few years, transistors Q2 and Q3 (IRF531 and 2N3904) begin to fail due to overheating, causing the supply to C3 to sag, resulting in one or more of a variety of symptoms, described in section 2.1.

Section 3.1.1 has already described how to replace the transistors, and increase heat dissipation. This section describes how to confirm the fault diagnosis and modify the circuit, allowing proper adjustment of the input to C3.

The computer board, underneath the right hand side of the control panel, is furthest from a direct feed from the +5V digital power rail, obtaining its power via the huge I/O board, which is mounted under the center of the control panel. You should therefore measure the voltage on the computer board, as well as the PSU, to check the +5V digital power rail.

Even though you can adjust the +5V supply by pre-set potentiometer R22, you will find that by the time the power has reached the computer board, the voltage has sagged. If the computer board falls below 4.89V the Chroma will begin to fail, and the power going into C3 must be increased. A computer board may function erratically as low as 4.84V, but any less than that will probably result in complete failure.

You should modify the PSU pre-regulator to allow proper control of C3.

A picture of the relevant section of the power board, with relevant notes, is given in the following diagram, as well in Addition to Pre-Rectifier Circuit and Addition to +12V and -12V Regulator below.

Power Supply Board Between Transformer and Giant Capacitor C3

Modifications for R46 adjustment potentiometer circuit: Replace R3 and R7 with flying leads going to new board holding R46 potentiometer adjustment resistor network circuit. Remove R44, R45 and any extra "hard adjustment" resistors. Flying lead connections are shown by F1, F2 and F3 in red.

Oscilloscope Trace of Chroma PSU from Terminals of Giant Capacitor C3

The Rhodes and ARP Chroma synthesizer power supply board has one very large vertically mounted capacitor, numbered C3, rated at 7.5 Volts, 47000uF, which indirectly powers all logic chips in the Chroma.

A common fault with aging of the pre-rectifier circuit, is that C3 is no longer sufficiently charged to power the logic chips, especially the computer board.

This diagram shows the oscilloscope trace across the terminals of C3, for a properly adjusted Chroma. The trace should oscillate between 6.0 and 7.4 Volts. The voltage from ground to the Computer Board 5 Volt rail should show at least 4.88 Volts, preferably 4.90 Volts. The voltage across the terminals of C3 should average between 6.30 and 6.33 Volts.

David Clarke [21030085++] points out that the Service Manual (Calibration and Checkout, third paragraph) suggests that the top voltage should not be greater than 7.0 volts.

You need to remove resistors R3 and R7 and replace then with wires going to a new circuit board, holding the new C3 input adjustment circuit. The new board will also hold the 11K resistor in series with a 2K potentiometer, used to adjust the +12V and -12V rails, as previously described in section (3.1.2).

R3 is the 22K resistor, to the right of the two big MR750 diodes CR1 and CR2, in front of the small 1N4742 diode CR3, to the left of 100 Ohm resistor R5 and 9.1K resistor R7.

R3 must be removed and replaced by a pair of wires going from the PSU to the new board. Power Supply Board Between Transformer and Giant Capacitor C3 and Addition to Pre-Rectifier Circuit show the connection nearest to the small 1N4742 diode CR3 as "F1". The R3 replacement connection furthest from the small 1N4742 diode CR3 is shown as "F2".

22K resistor R3 connects in series with 9.1K resistor R7, via the copper track on the circuit board, which points backwards, towards 100 Ohm resistor R5, which lies across the foot of the IRF531 power transistor Q2.

R7 is to the left of 33K resistor R6, behind the 2N2904 transistor Q3.

R7 must be removed and replaced by a single wire from the end furthest from R3. Diagrams for Power Supply Board Between Transformer and Giant Capacitor C3 and Addition to Pre-Rectifier Circuit show the connection as "F3".

Resistors R44 and R45 must also be removed.

R44 is 9.1K and is to the right of the pair of 1N4002 diodes CR6 and CR7, to the left of giant capacitor C3.

R45 is 15K and is in front of 2N2904 transistor Q3, to the left of giant capacitor C3, behind the pair of 1N4002 diodes CR6 and CR7.

Thus three wires replace R3 and R7, leading to the new circuit board, which holds the four resistors that have been removed from the PSU board, plus adjustment potentiometer R46.

Again, I must stress that potentiometer R46 CANNOT be soldered into position on the main PSU board, to the right of R45, because the copper tracks have changed somewhere between serial numbers 21000001 and 21010093, so soldering a potentiometer onto the board does NOT work.

Addition to Pre-Rectifier Circuit

See above diagram for connection points.

Only very early Chromas have this resistor adjustment network soldered into the position indicated by the schematic in the ARP Chroma Service Manual.

Most Chromas have a modified PSU board, with changes to the copper cladding connections which make it impossible to solder the adjustment components into the expected positions on the board.

Hence the need to remove R3, R7 and R45 and replace them with flying leads to a separate board.

Note that pre-set potentiometer R46 should be at least 0.2Watt, such as RS part number 186-772, which is 0.5Watt.

Addition to +12V and -12V Regulator

[See section 3.1.2 for connection to points]

11K 0.25 Watt 2% resistor in series with a 2K pre-set potentiometer to allow fine adjustment. Flying leads are used to make this a replacement for resistor R27.

When installing the modified board in the Chroma, set the initial +5V level with the R22 potentiometer on the main PSU board, then whilst measuring the computer board voltage, use R46 on the new board to adjust the input to C3, which will then store enough power to bring the computer board up to at least 4.89V and preferably more.

You should slowly adjust potentiometers R22 and R46 to give a safe balance between too high a voltage leaving the PSU and too low a voltage on the computer board. For best performance, R22 should be adjusted to show 5.05V from the yellow +5V analogue output wire and 5.04V from the orange +5V digital output wire from the PSU.

If you connect an oscilloscope and a digital multimeter across the terminals of C3, you should see the voltage cycling between a low point of 6.0 Volts and a peak of somewhere between 6.8 Volts and 7.4 Volts. The peak voltage depends on the sensitivity of the crowbar over-voltage protector, which varies considerably from Chroma to Chroma. The image under Oscilloscope Trace of Chroma PSU from Terminals of Giant Capacitor C3 above shows an image of the oscilloscope trace across the terminals of capacitor C3, for a properly adjusted Chroma, with a weak crowbar protector. A strong crowbar protector gives the same shape trace, with a lower peak value. Remember to calibrate your oscilloscope carefully, against a digital voltmeter, before attempting to adjust the power supply unit.

A digital voltmeter across the terminals of capacitor C3 should show an average of 5.8V for a strong crowbar over-voltage protector, or up to 6.33V for a weak protector.

Remember to test again after an hour, when the PSU is fully warmed up, also remembering to adjust both potentiometers R22 and R46, to achieve the best balance between over-powering the I/O board and under-powering the computer board. Also remember that the analogue circuits function best at 5.05V instead of exactly 5V.

If an external device is connected to the COMPUTER INTERFACE port on the back of the Chroma, you will need to perform a second adjustment with the extra device in position.

(3.1.4) Repairing the reset circuit

If the Chroma forgets program parameters after it has been switched off, try replacing the batteries.

If the Chroma still forgets parameters, the power supply board reset circuit is probably faulty. This controls switching on the CMOS RAM battery backup when mains power is switched off.

The reset circuit transistors Q12 (2N3906) and Q13 (2N3904) are prone to overheating and should be replaced, as previously described in section 3.1.1.

Diodes CR14, 15 and 16 are next most likely to fail, but CR14 is a 3.9V 1N5228B, which is hard to obtain, so try replacing the 1N4148 diodes CR15 and CR16 first.

If the Chroma forgets program parameters whilst it is still switched on, see section 4 for diagnosis and repair.

(4) Checking the Computer Board

The computer board, underneath the right hand side of the control panel, is furthest from a direct feed from the +5V digital power rail, obtaining its power via the huge I/O board, which is mounted under the center of the control panel.

Even though you can adjust the +5V level by pre-set potentiometer R22 on the power supply board, you will find that by the time the power has reached the computer board, the voltage has sagged. Remember that adjusting R22 to give a voltage any higher than 5.05V at the power board can damage TTL logic chips on the I/O board.

If the computer board falls below 4.89V the Chroma will begin to fail, producing symptoms described in section 2.1. A computer board may function erratically as low as 4.84V, but any less than that will probably result in complete failure.

You should therefore measure the voltage on the computer board, as well as at the power supply board outputs, to check that the +5V digital rail is delivering enough power. Anything less than 4.89V means that the power supply board pre-regulator circuit needs fixing, to allow proper control of reservoir capacitor C3, as described in section 3.1.3. It is highly recommended that you also read the rest of section 3, to perform other power supply board modifications at the same time.

Erratic behaviour can also be caused by bad connections between the I/O and computer boards, but this is comparatively rare.

If the Chroma forgets program parameters whilst it is still switched on, check that the CMOS RAM voltage is at least 4.89V. If the voltage is correct, try swapping the CMOS chips between sockets, to see if different parameters are forgotten, or replace them.

Remember that although the RAM chips all look the same, some are CMOS and some are static RAMs, so check the numbers carefully before swapping locations. The CMOS chips are 4334s, although 6514s would probably also work, and are held in sockets Z13 through Z18. The static RAMs are 2114s.

If the Chroma forgets program parameters after it has been switched off, try replacing the batteries. Contrary to the figure stated in the Chroma manual, if pressing SET SPLIT followed by 6 shows a CELL value of anything less than 3.00 volts, the CMOS chips may begin to loose their memory.

If the Chroma still forgets parameters, the power supply board reset circuit is probably faulty. This controls switching on the CMOS RAM battery backup when mains power is switched off. It should be repaired as described in section 3.1.4.

(5) Checking the EQ Board

If the Chroma responds to buttons, by changing the LEDs and DATA READOUT display, but produces no sound when the keyboard is played, you should check the EQ board and the mother board.

If you are using one of the "MONO OUT" jack sockets, swap to "AUDIO INPUTS/ OUTPUTS" jack number zero. If this works, the EQ board is faulty, otherwise the mother board has probably failed, in which case you should follow the instructions in section 6, to diagnose the problem.

The EQ board is found under the control panel, to the left of the huge I/O board. The 3360 Curtis VCA chips Z1 and Z2 are most likely to fail.

Please note that the outputs from the four "AUDIO INPUTS/OUTPUTS" jacks are clearer than the "MONO OUT" jack sockets, because the signal passes through fewer op-amps, so I recommend using them for recording music. They bypass the EQ controls, which are really only intended to adjust for deficiencies in PA systems at live performances.

(6) CHECKING THE MOTHER BOARD

If the Chroma responds to buttons, by changing the LEDs and DATA READOUT display, but produces no sound when the keyboard is played, you should check the EQ board and the mother board.

If you are using one of the "MONO OUT" jack sockets, swap to "AUDIO INPUTS/ OUTPUTS" jack number zero. If this works, the EQ board has failed. See section 5 for EQ diagnosis and repair.

If there is no signal from the default audio output socket number 0, try re- directing the output signal to socket numbers 2 or 3, which are driven by a different mother board op-amp (4558 Z19).

Re-direction is achieved by pressing the white "PARAM SELECT" button to put the Chroma into programming mode, then press the green button number 5, labelled "OUTPUT SELECT" button, then move the "PARAMETER CONTROL" slider until the "DATA READOUT" display shows "P5 2" or "P5 3".

If this works, either the first mother board op-amp 4588 Z18, or its controlling logic chip 4016 Z16, has failed, usually due to overheating.

Heat is a big problem, conducted onto the surface of several components by cables of short length, from the adjacent power supply board, often causing the following components to fail:

The Chroma digitally samples its own analogue output, comparing it with a sub-division of the system clock, for extremely accurate tuning. The instrument adjusts its own tuning control signals until perfection is achieved.

If the Sample & Hold section of any voice board fails, it can often kill one of the CMOS chips on the mother board. Chip Z8 is most likely to fail, which is a 4556 in the mother board's Sample & Hold Decoder.

After checking Z8, check Z4 and Z7, because these often fail.

This is best checked with a calibrated oscilloscope, since the logic output levels often sag, so that a logic probe shows that all is well, when the voltages are actually too low to work.

(7) Checking the Keyboard Scanner on the Stack Switch Board

If the keyboard scanner circuit fails, the entire Chroma may freeze.

If the Chroma appears to ignore the keyboard, try playing a very fast series of notes. If the two big program number LEDs flicker, the keyboard scanner circuit is working. If a MIDI interface is fitted, it should still output MIDI note information as normal, even if the Chroma is silent.

Check all logic chips with a calibrated oscilloscope, since the logic output levels often sag, so that a logic probe shows that all is well, when the voltages are actually too low to work.

(8) Checking the I/O Board

If all voices fail to tune, one possible cause is pre-set potentiometer R1 being wrongly adjusted. You must carefully adjust R1 until a digital voltmeter shows zero volts between the two adjacent test points. This is difficult to do, since this device should really have been implemented as a screw potentiometer. However, if all voices refuse to tune, it is usually a power supply board or mother board problem.

If all voices fail to tune, another possible cause is damage to capacitor C31, which appears in different positions, depending on the serial number of your Chroma.

The Chroma digitally samples its own analogue oscillator output, comparing it with a sub-division of the system clock, for extremely accurate tuning. The instrument adjusts its own tuning control signals until perfection is achieved.

Z10 is part of the timer circuit that compares a sub-division of the system clock, with the frequency of each analogue oscillator, when the voices are being tuned. Pin 11 should receive a clean 2MHz clock input. Any noise disrupts the tuning comparison, hence capacitor C31 is provided to keep the clock signal clean.

Timer circuit chip Z44 (74LS393) also uses the 2MHz clock on pin 1, which takes its feed from pin 11 of Z10.

On newer Chromas, C31 is the 50pF 20% capacitor connected to pin 11 of 74LS74 chip Z10, on the component side of the board, but on older Chromas, C31 is the only component mounted on the other side of the board, crudely soldered onto the back of pin 11 of Z10 or pin 1 of Z44.

You need to use an oscilloscope to monitor pin 11 of Z10 and pin 1 of Z44, whilst pressing SET SPLIT followed by 50, which will re-boot the Chroma's computer systems, causing all boards to be re-tuned.

AUTO TUNE will NOT help you, since it bypasses boards that have previously failed a tuning test. The system must be re-set to force it to examine all eight voice boards.

The other side of the tuning process is partly controlled by the mother board, where a zero-crossing detector should provide the I/O board with a clean clock signal, at the same frequency as the analogue oscillator being tuned.

Once again, any noise will disrupt the tuning process, so you need to use an oscilloscope to monitor the many pins of the timer circuit, where the zero- crossing detector signal "SYNTH ZCD" is applied. The circuit diagram is shown on page 6 - 5 of the Chroma hardware manual. C31 is an addition to the original design and is missing from the circuit diagram.

If the Chroma freezes with all LEDs illuminated, it can be caused by bad connections between the I/O board and the main computer board, though this comparatively rare. It is usually a power supply board problem.

If program parameter values appear to jitter between two adjacent values, it can be caused by a faulty ADC chip Z3 on the I/O board, but this is extremely rare. It is usually a power supply board problem.

The I/O board is only likely to fail because it holds many components. The weakest spots are all interfaces to other boards. At power-up, whilst capacitors are still charging, the voltage and current between boards can produce transient peaks, blowing logic chips such as:

Surprisingly, I never seen the heavily-loaded LED driver chips fail.

(9) Repairing Faulty Voice Boards

If all the voices fail to tune it can be caused by one voice board not being silent when the other boards are being tuned. Voice boards sometimes output noise due to a failure in the Z16 3360 VCA chip or the Z17 4052 multiplexer. The faulty board might alternatively be producing a signal due to faulty control voltage levels being applied to the Z16 3360 VCA.

If the power supply board appears to be failing, note that faulty voice boards often affect the power output. The CMOS chips are prone to fail with an internal short, or low resistance, causing the power rails to drop below their proper value.

If one of the power rails shows a low voltage, take out the voice boards one at a time to see if one board causes the instrument to fail. If removing a particular board cures the fault, replace all the other boards to check that it is not simply the quantity of boards being powered that the supply unit cannot cope with.

If an error occurs on one voice, or if a board fails to tune, it is usually the voice board itself that is at fault, though there are other possibilities as follows:

  1. The mother board is faulty
  2. The power supply is sagging and this is the first board to be affected
  3. The digital or analogue control signals are faulty at the either the computer board or the I/O board through which the signals travel.

Since the voice board itself is the most likely place to find the error, this discussion will start here.

If a voice board fails to tune, the DATA READOUT display will show something like this:

"Err    5"

In this example, voice board 5 has failed to tune. In extreme cases, you could see "Err 0357", or in the worst possible case "E01234567".

The Chroma has eight voice boards, numbered zero though 7. Each one is a complete dual-oscillator synthesizer, with truly remarkable capabilities, normally only found on huge modular systems such as the Roland System 700, the Moog 3C, or the ARP 2500.

The Chroma is actually the mighty ARP 2100, marketed as the Rhodes Chroma following ARP's demise.

Returning to the voice board problem, very few Chroma power supply units are strong enough to stabilise as soon the instrument is switched on. Most Chromas fail to tune a couple of boards at first, especially in particularly hot or cold weather.

The way around the problem is to wait 10 seconds, then press SET SPLIT followed by 50, which will re-boot the instrument's computer systems, causing all boards to be re-tuned.

AUTO TUNE will NOT help you, since it bypasses boards that have previously failed a tuning test. The system must be re-set to force it to examine all eight voice boards.

If this still doesn't cure the problem wait a minute or so and press SET SPLIT followed by 50 again. If the problem persists, the voice board numbered in the DATA READOUT display is probably genuinely defective.

If more than 3 or so boards fail to tune, the problem is probably not in the voice boards themselves, so sections 2 through 8 should be followed to identify the fault.

Having established that a voice board is truly faulty, the next stage is to press SET SPLIT followed by 31, to re-boot the instrument's computer systems in diagnostic and debugging mode.

This needs some explanation.

In debugging mode, all voices will sound in turn as keys are played, regardless of how defective they are.

If the "PATCH" program parameter is set to zero, 16 voices are available, 2 per voice board. These will play in the order of voice B first, then voice A (not A then B, as you might expect).

If the "PATCH" program parameter is set to 1-15, 8 voices are available, 1 per voice board, in dual oscillator mode.

The tuning algorithm tries to tune oscillator A first. If oscillator A fails to tune, the next action depends on the version of the operating system held in ROM. In some versions, oscillator B is set to a very low frequency and then bypassed, so that it does not track the keyboard. In other versions of the operating system, oscillator B is tuned one octave lower than oscillator A, and may or may not track the keyboard.

Use the SET SPLIT 31 debugging mode to understand how the oscillators are behaving. First listen to the voices as you play each subsequent note on the keyboard, then use an oscilloscope connected to test point rings 1 and 2 (TP1 and TP2), to view the waveforms prior to filtering.

After using the Chroma's debugging mode to reveal the oscillators behaviour, you can switch the instrument off and take out the defective board immediately, because 90 percent of voice board faults are caused by the three CMOS logic chips that control the sample and hold bank. These are the 4051 8-way multiplexers, numbered Z19 and Z20, plus the 4556 dual 1 of 4 decoder, numbered Z25.

I strongly recommend that you don't even start tracing through the voice card circuits for any fault until you have replaced all three of Z19, Z20 and Z25, to see if they cure the problem. Remember to use proper mounting sockets, because you can expect to replace these chips at regular intervals of anything between 1 and 10 years. Some boards seem to load the chips more heavily than others. I don't know why.

If you have replaced Z19, Z20 and Z25, but the fault remains, don't worry because you would have to replace them eventually anyway. Isolating the fault is easy, now that the sample and hold logic chips are known to be working properly.

When using a voice program with a standard VCO-VCF-VCA patch, op-amp Z10 amplifies the oscillators and feeds the signal to the filters, then op-amp Z14 amplifies the filter output and feeds it to the voltage controlled amplifiers.

Thus, displaying the outputs from Z10 and Z14 on an oscilloscope will isolate the fault to either the oscillator, or filter, or amplifier circuit.

For proper testing, you must enter a voice program into the Chroma that puts the synthesiser into 16-voice mode, outputting a square-wave signal from each voice. This ensures that a standard VCO-VCF-VCA configuration is being used and makes any distortion of the waveform easily visible on an oscilloscope.

Entering the test program is easy, because the Chroma contains a default test program in ROM, which is very close to the one we need.

Switch on the Chroma and see if the yellow LED above the PARAM SELECT button is on. If not, press PARAM SELECT to put the instrument into programming mode, which will switch the LED on.

Next, look at the green and red LEDs above the EDIT A and EDIT B buttons. You need to have both of these switched on, so you have to press both EDIT A and EDIT B simultaneously. Pressing only one button will toggle between A and B, but we need both LEDs switched on, to allow the two voice channels on each board to be programmed at the same time.

Next, use your left hand to press the PARAM SELECT button. Keeping PARAM SELECT pressed, with the yellow LED on, use your right hand to press every PROGRAM SELECT/PARAMETER SELECT button from 1 though to 50. This will retrieve a default voice program parameter value from ROM, for each of the 100 parameters (A 1-50 and B 1-50) that make up a voice program.

If you think you missed out a button, or didn't press PARAM SELECT hard enough, just repeat the process. You can do this as many times as you like.

The default voice program is very close to the one we need, except that the waveform and pulse width values are wrong.

Press the WIDTH button, which is number 34, then use the PARAMETER CONTROL sliding potentiometer to set the DATA READOUT display to show "P34 32".

Then press the WAVE SHAPE button, which is number 33, and use the PARAMETER CONTROL sliding potentiometer to set the DATA READOUT display to show "P33 1".

The resulting voice program is held in volatile RAM and will be forgotten when you switch off the instrument, unless you deliberately store it in one of the Chroma's 50 voice program memories. It should give an unmodulated single oscillator square wave output from each of the 16 voices (2 voices per board in 16-channel mode).

Now you can use an oscilloscope to look at the outputs from the 4558 op amp number Z10. Pin 1 shows oscillator A output and pin 7 shows oscillator B.

If the waveform is not a perfect square wave, the fault lies in the oscillator circuit, which you should trace through in detail.

If the waveform is correct, look at the outputs from the TL082 op amp number Z14. Pin 1 shows filter A output, pin 7 shows filter B.

If the waveform is not a very nearly perfect square wave, the fault lies in the filter circuit, which you need to trace though in detail. If the wave-form looks good, then the fault lies in the VCA circuit, which you need to trace through in detail.

This should be all the information you need to fix a Chroma.