Trying to fix power supply on Soundcraft GB 8 40 channel mixing board

Thread Starter

nbtone

Joined Oct 14, 2016
65
I see where it says live ground with a white arrow head. I see where it says DGND with a white arrow head. I see several black arrow heads , and I see several slashes under a line that looks like a ground of some kind.
Are the voltages past the bridge rectifier DC? I measure 163 VDC across C34.
The chip is available here for $11: https://americas.rsdelivers.com/pro...3V-rvIJ5Qwu0tzqOIyUspubQex5HxKMBoCcWUQAvD_BwE


The green wire attaches to chassis ground. Live ground is not actually a ground, it is a floating reference point for the line connected supply. Transformer TX1 separates the line (hot) side of the unit from the low voltage side on the right of the schematic. Everything to the left of TX1 is deadly, use extreme caution. The right side is safer but the power supply puts out 48 volts and this is still in the range of deadly voltages. (i.e. voltages that can supply sufficient current to kill you. It is the current that kills and it only takes 100 to 200 milliamps.)

Without an oscilloscope you will have a had time troubleshooting this power supply. This is a switch mode power supply and a meter will not give much information. However, there are a couple of things you can check if you do so safely.

Here is a brief description of how the circuit works and what can be tested with a voltmeter:

Of course you know the purpose of the bridge rectifier and that you can measure the high voltage DC across C34. Since you have 116 volts input, the voltage across C34 should be 116 x 1.41 = 164 volts. (RMS x the square root of 2) The negative side of C34 is "live" ground. If you had a scope you would see it with a negative half cycle waveform at the line frequency. So "live" ground is very hot! The positive side of C34 is labeled +V1 and is also very hot.
If you have a good DC voltage of about 160 to 170 volts across C34 then you can proceed to the next step. IC3 is the heart of the power supply, it was manufactured by a company called Unitrode. Unitrode was bought by Texas Instruments and this particular part is no longer made. Here is how it works:
At start up a small amount of current flows through R24 to supply IC3. When IC3 gets sufficient power it will start oscillation on pin 4 which connects to the gate of TR2. TR2 supplies a little current through the primary winding of TX1. The secondary winding on terminals 3 and 14 supplies some small current through R30, D10 to TR4. TR4 is a linear regulator, the base is set to 13 volts (relative to live gnd) by R31 and zener diode ZD3. Due to the voltage drops across the base-emitter of TR4 and D11, the voltage supplied to pin 3 of IC3 will be about 11 volts. (referenced to live gnd) So the result is that R24 gives it a little kick but the actual voltage supply to IC3 is coming from the switching of the power circuit.
With a multimeter you can measure some of the DC voltages but you won't be able to see the actual switching. If the voltages on the primary side appear OK then check the secondaries, there are four outputs; +17, -17, 14 and 48 volts. In fact, check these first, it is a lot safer.
 

drc_567

Joined Dec 29, 2008
1,156
... C34 looks ok to that point. Maybe try to verify some of the other DC voltages as suggested previously.
Check the 17 volt DC voltage source. This seems to be the one that powers the two LED diodes. The schematic shows a kind of terminal strip with several DC voltages, but I'm not sure where it is actually located.
This is one of those complex circuits, and may be difficult to troubleshoot.
 
Last edited:
The ground with slashes is the chassis ground, it should have the green wire from the line input attached to it. Toward the lower right you will find this diagram:
upload_2019-5-23_20-16-35.png
This shows DGND and the black triangle connected to chassis ground through a couple of links. DGND is probably the mixer circuit ground. I'm not certain what the black one is for, it may be for filtering some noise.
The 'live" ground is a reference point for all of the high voltage circuit.Any point in the high voltage side needs to be referenced from it.Unfortunately, it has high voltage on it with respect to the chassis ground. You can check this by using your multimeter to measure the AC voltage between those two points. It will probably be close to 60 volts RMS.
If you are careful, you can place the negative multimeter lead on live ground and measure the DC voltage at various points on the high voltage side. I say to be careful because there is a very good change of getting electrocuted.
I would invest in a GFCI outlet. A GFCI outlet will interrupt the line voltage if current flows to ground. It might save your life. I don't know where you live but I can get one locally for less than $15. It is a very small price to pay for better safety. However, you need to have it installed by an electrician. You may be able to find a portable one that would not require installation.
 
I checked your source and they are out of stock but Digikey has 49 in stock: https://www.digikey.com/products/en?mpart=UCC35702N&v=296
If you are located in the U.S. you can get it easily, it not you will have to find a TI distributor.
I also found it on the TI website and they have the specification sheet. (attached)

Lack of an oscilloscope still hinders your ability to test this. Ask around locally, you may find someone to assist.
 

Attachments

JoeJester

Joined Apr 26, 2005
4,390
You have the live ground in the primary side of the transformer. This is a floating return to the transformer.

You have earth ground at the AC input which is tied to neutral back at the electrical distribution.

You have Dgnd (white triangle) and the black triangle tied together via Lk1 And Lk2 to a point I assume is the chassis grounds ... As the angles are more than the normal chassis ground symbol.

This could confuse the person working on a system, especially if their skills are less than optimal.
 

Thread Starter

nbtone

Joined Oct 14, 2016
65
I have connected the earth ground to the correct screw connection point on the circuit board. Now there is continuity from earth ground to dgnd and the black arrow.
I measured the voltage at pin 3 of the chip and found about 9.5 VDC, using the live ground at the negative terminal of c34 for ground reference.
I measured the voltages at ch2 and found 0 bolts where there is supposed to be + or - 17v, and I found 365mv where the 48 v is supposed to be. I used the earth ground, which is connected to the black and white arrows, as the ground reference for this test.
If I find someone with an oscilloscope what would I need to do to find which parts need replaced?
 

drc_567

Joined Dec 29, 2008
1,156
It may be that the TX1 transformer is not receiving a pulse waveform on its primary side for one reason or another. With a scope, you would be able to examine pin 4 of IC3, the UCC37502 chip, in order to see if things are working up to that point.
... Again, there are big problems with connecting a scope to a live AC circuit. Usually, plugging the scope into an isolation transformer is required. A GFCI outlet may work, but get another opinion.
... There are 3 or 4 parts besides IC3 that may be where the problem is.
 
Last edited:

JoeJester

Joined Apr 26, 2005
4,390
There are only three ways to isolate an oscilliscope circuit. The easiest is a DC operated oscilloscope and they are pretty cheap for a single channel hobbyist model with limited bandwidth. Those scopes do not come with x10 probes so you would end up spending a little more money for that item. The few I saw at Amazon for under $50 were 200 kHz bandwidth.

Most scopes are earth grounded via the green wire. The best way is to use an isolation transformer. I've heard of people just removing the earth ground connection. I've never tried using that because I didn't want to do the paperwork if someone killed themselves.
 
Last edited:

mlsirkis

Joined Aug 11, 2010
32
I see where it says live ground with a white arrow head. I see where it says DGND with a white arrow head. I see several black arrow heads , and I see several slashes under a line that looks like a ground of some kind.
Are the voltages past the bridge rectifier DC? I measure 163 VDC across C34.
You may want to consider disconnecting the wires that connect the output of the supply to the rest of the console and adding resistors as a dummy load. This serves two purposes. First, it will prevent damaging the rest of the console if you accidentally slip while probing the power supply. Second, it’s possible that there is a short somewhere in the console that is pulling the supply down.
 

Thread Starter

nbtone

Joined Oct 14, 2016
65
You may want to consider disconnecting the wires that connect the output of the supply to the rest of the console and adding resistors as a dummy load. This serves two purposes. First, it will prevent damaging the rest of the console if you accidentally slip while probing the power supply. Second, it’s possible that there is a short somewhere in the console that is pulling the supply down.
I’m sorry. I never said whether I had disconnected the rest of the console before I tested the power supply. Would that make a difference? I have the power supply all the way out of the console and sitting on a table. All the testing was done on the table with the earth ground screwed to one of the circuit board mounting holes.
 

Thread Starter

nbtone

Joined Oct 14, 2016
65
It may be that the TX1 transformer is not receiving a pulse waveform on its primary side for one reason or another. With a scope, you would be able to examine pin 4 of IC3, the UCC37502 chip, in order to see if things are working up to that point.
... Again, there are big problems with connecting a scope to a live AC circuit. Usually, plugging the scope into an isolation transformer is required. A GFCI outlet may work, but get another opinion.
... There are 3 or 4 parts besides IC3 that may be where the problem is.
I see the need for an isolation transformer. Thank you! Electrocution and damage to the oscilloscope are avoided with the use of one. Here is the YouTube video about it is hat I watched : https://youtu.be/LbkCU-LBqKk
 
I've been on vacation and am just catching up.
An isolation transformer will permit you to connect live ground to chassis ground and make direct measurements. However, using the transformer removes the protection of the GFCI and is less safe. You can still use an oscilloscope to make measurements without the transformer if it is dual channel. Set channel 2 to inverted and "add" the channels. When making measurement place the channel 2 probe on live ground and use the channel 1 probe for the point to be measured. This is called differential mode operation, here is a video explaining how to use it:

A 200 KHz oscilloscope is inadequate for this circuit. Note on the IC spec sheet that it can operate up to 700 KHz. You will need at least 1MHz bandwidth.
 

Thread Starter

nbtone

Joined Oct 14, 2016
65
I've been on vacation and am just catching up.
An isolation transformer will permit you to connect live ground to chassis ground and make direct measurements. However, using the transformer removes the protection of the GFCI and is less safe. You can still use an oscilloscope to make measurements without the transformer if it is dual channel. Set channel 2 to inverted and "add" the channels. When making measurement place the channel 2 probe on live ground and use the channel 1 probe for the point to be measured. This is called differential mode operation, here is a video explaining how to use it:

A 200 KHz oscilloscope is inadequate for this circuit. Note on the IC spec sheet that it can operate up to 700 KHz. You will need at least 1MHz bandwidth.
I'm looking at this one: https://www.picotech.com/oscilloscope/2000/picoscope-2000-overview
Do you think it would work? I'm thinking that it might help with testing the rest of the board after the power supply is fixed with all the other functions that it has.
 
I used a PicoScope several years ago and was generally satisfied with the performance. But I used it strictly for low voltages. It has a +-20 volt input limit. Even with a 10 to 1 probe it would barely be adequate for your high voltage needs. It would probably also have some distortion in the waveform at elevated voltages. Also be aware that it will be connected to your computer and could damage it if you connect it incorrectly.
So, good for low voltage, general purpose electronics but unknown for high voltage. Check online, see if any reviews or other information will help your decision.
 

Thread Starter

nbtone

Joined Oct 14, 2016
65
I’m looking at two, and I’m not sure which would be better:

Bk Precision 2520 Digital Storage 20mhz Variable Sweep Oscilloscope
Or
Hantech DSO 5102P

The first one says 400v max, and I don’t see a voltage rating on the Hantech
 
I have two Hantek oscopes, a 6022BL USB scope and a DSO8060. I have never used the 6022 to observe high voltage but I have looked at line voltage using the DSO8060. The 8060 has the vertical range but the waveform was a little distorted but is usable.

Looks like the BK is only available used and has a real CRT instead of an LCD screen. Personally, I like the Hantek better, it is lighter and more portable. The input can stand 300V at low frequency with a 10:1 probe. Above 100KHz the max level is derated at 20dB per decade down to 13volts at 3MHz and stays at that level to full bandwidth. Still not a bad scope but good for the price.
 

Thread Starter

nbtone

Joined Oct 14, 2016
65
I have two Hantek oscopes, a 6022BL USB scope and a DSO8060. I have never used the 6022 to observe high voltage but I have looked at line voltage using the DSO8060. The 8060 has the vertical range but the waveform was a little distorted but is usable.

Looks like the BK is only available used and has a real CRT instead of an LCD screen. Personally, I like the Hantek better, it is lighter and more portable. The input can stand 300V at low frequency with a 10:1 probe. Above 100KHz the max level is derated at 20dB per decade down to 13volts at 3MHz and stays at that level to full bandwidth. Still not a bad scope but good for the price.

I found a BK that has never been used and still in the box. I wouldn’t need 10x probes to use it, would I? The Hantech has more features, but I don’t know if I will need them.
 
Yes, you will need 10x probes but any probe you get should have a switch to set 10x. You at least need the ability to set differential mode on the scope. Differential mode is achieved by inverting one channel and adding the channels. This enables you to view signals when the reference point is floating with respect to ground. You use channel 2 as the reference and make the measurement with channel 1. You will need two probes.

I would really like to discourage you from attempting to repair the sound board. It is too dangerous for a beginner to work on line connected devices. Despite my earlier warnings you appear intent to continue. That power supply is death waiting to happen. I won't post anymore on this topic. If you have questions on other topics, I'll be glad to help.
 

Thread Starter

nbtone

Joined Oct 14, 2016
65
I should have expected as much. I’m in the business of audio electronics, because I discovered a way to make almost any bass amplifier to sound a whole lot better, therefore making it possible to make for a much more listenable experience at a live performance venue. My power supply is not a switch mode power supply, and I don’t believe a switch mode power supply would work in the circuit.
I’m used to this kind of treatment really, so I’m not surprised. I’ll figure it out some other way. I’ve been doing this kind of stuff for years without a scope, but now I think I’ll learn how to use one for this project.
Thanks for your advice, otherwise.
 
Top