Well I thought I had it fixed but when I mounted it up to the outputs, It won't adjust properly. With the light bulb in line it won't go higher than 18V rails. If I disconnect the output wires on the board then it will go to full rail. I'm reading about 1.2V & -1.2V on the outputs of the PCB when not connected to the output transistors. As soon as I connect them back up, the light bulb burns bright and the rail drops to 18V again and the drivers drop to .5V.
This is all with the bias pot set to minimum. Any ideas? What voltage should I be seeing at the outputs of the board?
Another thought. I'm using transistor sockets for my outputs. Could they cause that?
I was able to adjust the off set on the output and it stayed steady throughout the test. For some reason the outputs are reacting like they are over driven.
Any ideas wil be greatly appreciated.
Thanks, Terry
This is all with the bias pot set to minimum. Any ideas? What voltage should I be seeing at the outputs of the board?
Another thought. I'm using transistor sockets for my outputs. Could they cause that?
I was able to adjust the off set on the output and it stayed steady throughout the test. For some reason the outputs are reacting like they are over driven.
Any ideas wil be greatly appreciated.
Thanks, Terry
Hi Terry,
+-1.2v from the drivers is too high. You need to reduce this BEFORE you power up the output stage.
I have not built ksa, but from the schematic you need one Vbe drop for each half of the output to turn on the Iq and then you add on the Vdrop of the Re (about 200mV) so total AFTER you have proved everything is OK is about 850mV. NO it's not, are you running 12 output pairs?
Total Iq =1.9A, Iq per output pair = 158mA, VRe = 107mV.
Your total is really 757mV. But you do NOT set this.
You set minimum driver output voltage, then connect the output stage thro' the light bulb, then bias the output stage until the rail volts just starts to drop. Then stop and check the V drop on each Re, all should be within 10%. Now remove the light bulb ( a switch would be better to short out the bulb because this would give a soft start).
As suggested earlier leave the Pre setup Iq and check ALL your semi temps. Then start to ramp up the Iq in stages stopping to check Re V drop and temps. This will prove your heatsink capacity as well.
Terry, are you still awake? This is important if you want to stop resoldering replacement components!!!!
Finally do not blame the Ebay system when you buy equipment that is advertised for spares only and find it works (not) as described.
+-1.2v from the drivers is too high. You need to reduce this BEFORE you power up the output stage.
I have not built ksa, but from the schematic you need one Vbe drop for each half of the output to turn on the Iq and then you add on the Vdrop of the Re (about 200mV) so total AFTER you have proved everything is OK is about 850mV. NO it's not, are you running 12 output pairs?
Total Iq =1.9A, Iq per output pair = 158mA, VRe = 107mV.
Your total is really 757mV. But you do NOT set this.
You set minimum driver output voltage, then connect the output stage thro' the light bulb, then bias the output stage until the rail volts just starts to drop. Then stop and check the V drop on each Re, all should be within 10%. Now remove the light bulb ( a switch would be better to short out the bulb because this would give a soft start).
As suggested earlier leave the Pre setup Iq and check ALL your semi temps. Then start to ramp up the Iq in stages stopping to check Re V drop and temps. This will prove your heatsink capacity as well.
Terry, are you still awake? This is important if you want to stop resoldering replacement components!!!!
Finally do not blame the Ebay system when you buy equipment that is advertised for spares only and find it works (not) as described.
Mark A. Gulbrandsen said:
R108 feeds the Zener. In the original schematic Jan made it 1K58, on the BOM it has a 1K5/0.5 watt rating.
At 0.5 watts R108 will go up in smoke when the current through it exceeds 18.3 mA.
Suppose the entire 18.3 mA goes through the Zener diode, nothing through R110.
Voltagedrop over the zener is 27 volts, 27 times 18.3mA is ~0.5 watt.
Voltagedrop from the 27 volts at zener zd101 to the -44V negative rails is a bit over 70 volts. Total resistance of R110 to R119 is around 10k, so current through it will be around 7 mA.
At 17volts over R108, current through it is like 11.3 mA.
With 7 mA going in R110 average current into ZD101 is 4.3mA.
27 volts times 4.3 mA is ~ 0.12 watts dissipation.
Zener diodes always run pretty hot, with 44 volts rails even a 0.3 watt zener would do.
A 1watt zener diode can take more current than R108 would survive.
I suppose the 3watt zener for ZD101 on Jan Dupont's BOM is a typo.
Give me boards, i want to be a machinist too.
An amplifier with no bias at zero input is a class B amplifier.
With a small bias to overcome crossover distortion it becomes a class AB amplifier, bias at say 100mA.
If the non-linear characteristic of complementary devices are perfectly symmetrical these will cancel eachother out with the proper bias setting and they'll slide in class B without crossover distortion.
With bias set so high that up till clippingpoint none of the output devices close it is a full class A amplifier.
Anything inbetween might be called a high bias class AB amplifier.
In a sense the original KSA50 was a high bias class AB amplifier, not full class A, as clipping occured at over 75 watts output.
Vbe(on) for MJL21193/4 = 2.2 volts
Vbe(on) for MJL15003/4 = 2.0 volts
Below these values on the base these devices don't open.
Only small signal devices like BC550-BC414 and the like have a Vbe(on) of 0.6v.
At 1.8 amp bias for 3 devices in parallel class A output is 50w/8.
With 0.68 Ohm emitter resistors voltagedrop over the 0.68 ohm is ~0.4 volts for a bias of 0.6 amp per device.
Using MJL21193/4 voltage on the base needs to be 2.6 volts.
For MJL15003/4, its 2.4 volts.
Vbe(on) for the 2SC3955 is 1 volt.
With the trimmpot R126 set at maximum,5K, current through R125 is 1/5820= 0.17 mA.
Voltage over R124= 4700*0.17/1000= 0.8volts
So the minimum voltage between the bases of the + and - outputs is 1.8 volts.
Which means in the maximum resistance setting of the trimmpot voltage on the outputs should be at +/- 0.9 volts.
For 1.8 amp bias setting using MJL21193/4:
1 volt on the base of 2SC3955 divided by (X + 820) = i
5.2 divided by i = 4700 + 820 + X
=>> resistance value of the trimmpot = X ~ 300 Ohms
=>> i is current through R125 and the trimmpot ~ 0.9 mA
(5.2 = 2 times 2.6 volts on the bases of + and - output)
For 1.8 amp bias setting using MJL15003/4:
1 divided (X + 820 ) = i
4.8 divided by i = 4700 + 820 + X
=>> resistance value of the trimmpot ~ 417 Ohms
=>> i ~ 0.8 mA
Vbe(on) for MJL15003/4 = 2.0 volts
Below these values on the base these devices don't open.
Only small signal devices like BC550-BC414 and the like have a Vbe(on) of 0.6v.
At 1.8 amp bias for 3 devices in parallel class A output is 50w/8.
With 0.68 Ohm emitter resistors voltagedrop over the 0.68 ohm is ~0.4 volts for a bias of 0.6 amp per device.
Using MJL21193/4 voltage on the base needs to be 2.6 volts.
For MJL15003/4, its 2.4 volts.
Vbe(on) for the 2SC3955 is 1 volt.
With the trimmpot R126 set at maximum,5K, current through R125 is 1/5820= 0.17 mA.
Voltage over R124= 4700*0.17/1000= 0.8volts
So the minimum voltage between the bases of the + and - outputs is 1.8 volts.
Which means in the maximum resistance setting of the trimmpot voltage on the outputs should be at +/- 0.9 volts.
For 1.8 amp bias setting using MJL21193/4:
1 volt on the base of 2SC3955 divided by (X + 820) = i
5.2 divided by i = 4700 + 820 + X
=>> resistance value of the trimmpot = X ~ 300 Ohms
=>> i is current through R125 and the trimmpot ~ 0.9 mA
(5.2 = 2 times 2.6 volts on the bases of + and - output)
For 1.8 amp bias setting using MJL15003/4:
1 divided (X + 820 ) = i
4.8 divided by i = 4700 + 820 + X
=>> resistance value of the trimmpot ~ 417 Ohms
=>> i ~ 0.8 mA
My dad used to be Napoleon !
Just trying to clarify to Terry that if he is using Jan's circuit there must be something wrong with his boards if he is turning his trimmpot to minimum resistance and measuring +/- 1.2 volts on the outputs.
In the original schematic, when setting the trimmpot to zero resistance the voltage difference over the outputs would be 6.73 volts.
btw:
NUTTR posted at the time he is using 3 watt Zener diodes cause they were the only he had. Likely the reason for this number to be present at the BOM of the KSA50 Jan made.
Just trying to clarify to Terry that if he is using Jan's circuit there must be something wrong with his boards if he is turning his trimmpot to minimum resistance and measuring +/- 1.2 volts on the outputs.
In the original schematic, when setting the trimmpot to zero resistance the voltage difference over the outputs would be 6.73 volts.
btw:
NUTTR posted at the time he is using 3 watt Zener diodes cause they were the only he had. Likely the reason for this number to be present at the BOM of the KSA50 Jan made.
Is it just me? For some reason I can not get Babblefish to translate the last 10 pages into plain English!
Surely it can not be anywhere near as complex to wang around a potentiometer and measure the voltage drop acoss a resistor as the last 10 pages would make it seem?
If I was in Terry's position I am not sure if I could decipher the techno babble spewed here and I am of many things an Electronics Technologist.
Does anyone on this thread no how to post a straight forward technical answer without trying to prove how smart they are at every opportunity?
Anthony
Surely it can not be anywhere near as complex to wang around a potentiometer and measure the voltage drop acoss a resistor as the last 10 pages would make it seem?
If I was in Terry's position I am not sure if I could decipher the techno babble spewed here and I am of many things an Electronics Technologist.
Does anyone on this thread no how to post a straight forward technical answer without trying to prove how smart they are at every opportunity?
Anthony
Anthony,
Thats exactly what I've been thinking the last few days myself. Thats one reason why I always keep my posts towards the simpleton method.
Most in this thread are not the engeneer type...... heck, I still can't spell engeneer let alone be one...... I wouldn't want to be anyway.... I'd much rather build amps!
Mark
Thats exactly what I've been thinking the last few days myself. Thats one reason why I always keep my posts towards the simpleton method.
Most in this thread are not the engeneer type...... heck, I still can't spell engeneer let alone be one...... I wouldn't want to be anyway.... I'd much rather build amps!
Mark
Hi guys,
First I want to say how awesome it is that everyone is so willing to help on this forum. I really appreciate jacco's willingness to lay everything out and do all the math. Most of it goes over my head but it usually stops all the guessing as to what "might" be.
Now. I know there is something wrong with my boards. Otherwise I would be sing praises and listening to music.
It may be that I used slightly different values for some of the resistors. I tried to get as close as I could, but I wasn't able to find exact matches for everything. I wrote down the ones that were different but I have misplaced the list. When I went through and measured all of the resistors as I was looking for differences between the board, they seemed to be in the range that they were supposed to be.
I know it would take a little bit of work, but if someone who has a working model, could take some voltage measurements of each transistor as related to ground and make a list, it would help immensely when trying to troubleshoot. I'm sure if I posted where my board deviated for the correct voltage, someone would be able to tell me what might be causing it. I could post a list from mine but I think a list with the correct voltages might save confusion in the long run. The list could be added to the build wiki.
Through all of the last few posts, I still don't know what voltage I should be seeing at the outputs at the minimum setting. I guess even if I did I may not know where along the circuit it went astray.
Thanks, Terry
First I want to say how awesome it is that everyone is so willing to help on this forum. I really appreciate jacco's willingness to lay everything out and do all the math. Most of it goes over my head but it usually stops all the guessing as to what "might" be.
Now. I know there is something wrong with my boards. Otherwise I would be sing praises and listening to music.
It may be that I used slightly different values for some of the resistors. I tried to get as close as I could, but I wasn't able to find exact matches for everything. I wrote down the ones that were different but I have misplaced the list. When I went through and measured all of the resistors as I was looking for differences between the board, they seemed to be in the range that they were supposed to be.
I know it would take a little bit of work, but if someone who has a working model, could take some voltage measurements of each transistor as related to ground and make a list, it would help immensely when trying to troubleshoot. I'm sure if I posted where my board deviated for the correct voltage, someone would be able to tell me what might be causing it. I could post a list from mine but I think a list with the correct voltages might save confusion in the long run. The list could be added to the build wiki.
Through all of the last few posts, I still don't know what voltage I should be seeing at the outputs at the minimum setting. I guess even if I did I may not know where along the circuit it went astray.
Thanks, Terry
Terry,
I can certainly do this but it would be refrenced to +/- 38 volts dc rails. I don't know if that would do you any good or not since your rails are higher than mine but let me know as I'd be happy to do that. Any chance that you may have burned a trace on the board? I heve extra " JAN Boards" if you find that you have a bad one. Owing to the possible curents that can pass through some of the taces of things are mis wired this may be possible. Also is the bias device good and is the mounting tab FOR SURE isolated from the case? It ought to be but you never know.......
THE BOARD WIKI IS NOW CLOSED!
Mark
I can certainly do this but it would be refrenced to +/- 38 volts dc rails. I don't know if that would do you any good or not since your rails are higher than mine but let me know as I'd be happy to do that. Any chance that you may have burned a trace on the board? I heve extra " JAN Boards" if you find that you have a bad one. Owing to the possible curents that can pass through some of the taces of things are mis wired this may be possible. Also is the bias device good and is the mounting tab FOR SURE isolated from the case? It ought to be but you never know.......
THE BOARD WIKI IS NOW CLOSED!
Mark
THE BOARD WIKI IS NOW CLOSED!
Anyone still desiring to add boards or who wants to sign up... its now too late! Any left over baords will be available through the KSA-50 web site AFTER everyone has received the boards they have signed up and paid for.....
You will receive an e-mail from me about this over the next week while I am on the road. YOU need to calculate how much postage and shipment you desire to add to the total price...... I willi ndicate the weight in the e-mail per board. Over payments will not be returned but will donated to help fund this site. Payments will be accepted by Pay-Pal or direct mail if you don't have Pay-Pal. I have not decided if I will charge a 3% Pay-Pay surcharge but I am leaning towards doing that so I don't get short changed myself. I will be making ZERO profit from this.
ALL PAYMENTS MUXT BE IN U.S. DOLLARS.
Thanks to everyone for making this a very sucessful WIKI!
Mark
KSA-50 Web Site
Anyone still desiring to add boards or who wants to sign up... its now too late! Any left over baords will be available through the KSA-50 web site AFTER everyone has received the boards they have signed up and paid for.....
You will receive an e-mail from me about this over the next week while I am on the road. YOU need to calculate how much postage and shipment you desire to add to the total price...... I willi ndicate the weight in the e-mail per board. Over payments will not be returned but will donated to help fund this site. Payments will be accepted by Pay-Pal or direct mail if you don't have Pay-Pal. I have not decided if I will charge a 3% Pay-Pay surcharge but I am leaning towards doing that so I don't get short changed myself. I will be making ZERO profit from this.
ALL PAYMENTS MUXT BE IN U.S. DOLLARS.
Thanks to everyone for making this a very sucessful WIKI!
Mark
KSA-50 Web Site
Terry,
I know that you have gone through the boards many times. But there are usually easy things that might have been overlooked. Let me give a list of what things I would look for if I was in your situation:
- Check PSU connections
- Check ground returns. All grounds should be connected to that star ground including the common that is formed with the PSU.
- Check polarities of all transistors. Remember that orientation of different transistor may have leads which are not the same as what the original design is calling for.
- Check for cold solder with a magnifying glass and redo when necessary.
And as said by someone before, check the main board first and then continue with drivers and output.
I know that you have gone through the boards many times. But there are usually easy things that might have been overlooked. Let me give a list of what things I would look for if I was in your situation:
- Check PSU connections
- Check ground returns. All grounds should be connected to that star ground including the common that is formed with the PSU.
- Check polarities of all transistors. Remember that orientation of different transistor may have leads which are not the same as what the original design is calling for.
- Check for cold solder with a magnifying glass and redo when necessary.
And as said by someone before, check the main board first and then continue with drivers and output.