I recently built an F5 from diyAudio boards. All parts from Mouser, except input JFETs, bought, as suggested from Spencer. In diyAudio 4U case.
Amp powers up fine, bias is set to .59v with virtually no offset at the output. Music comes though, sounds nice. One problem.... output on right channel is low. Put in 2 volts 1K hz signal from old HP generator. Measured output on a scope. Left channel is ~12vAC out, right channel is about ~6.5 vAC out. So, the gain in the right channel is roughly half. Output FETs seem to have correct voltages, bias set correctly. Only change from original plans is that I am using a 16v transformer vs. 18v as specified. 18v was out of stock, and I had access to a 16v 300v a unit. Output of power supply is roughly +/- 20vDC, using diyAudio PSU.
I am fairly new at FET troubleshooting, pretty good with tubes. Ideas? I am thinking the input FETs are bad/weak. I did order an extra set from Spencer, but want to check other things before replacing.
Thanks for any advice.
Amp powers up fine, bias is set to .59v with virtually no offset at the output. Music comes though, sounds nice. One problem.... output on right channel is low. Put in 2 volts 1K hz signal from old HP generator. Measured output on a scope. Left channel is ~12vAC out, right channel is about ~6.5 vAC out. So, the gain in the right channel is roughly half. Output FETs seem to have correct voltages, bias set correctly. Only change from original plans is that I am using a 16v transformer vs. 18v as specified. 18v was out of stock, and I had access to a 16v 300v a unit. Output of power supply is roughly +/- 20vDC, using diyAudio PSU.
I am fairly new at FET troubleshooting, pretty good with tubes. Ideas? I am thinking the input FETs are bad/weak. I did order an extra set from Spencer, but want to check other things before replacing.
Thanks for any advice.
On a tube amp you would start doing a voltage table, measuring the voltages on all the active devices, without any signal fed into the amp.
I suppose the same would help here. If you compare the good side with the bad side, you will probably quickly be able to dial in to the area were the problem is.
I havn´t fixed mine yet. Just got the power supplies yesterday.
Best regards
Arthur.
I suppose the same would help here. If you compare the good side with the bad side, you will probably quickly be able to dial in to the area were the problem is.
I havn´t fixed mine yet. Just got the power supplies yesterday.
Best regards
Arthur.
Lot's of commercial gear is not earthed nowadays, but in the EU there are some rules that apply to non-earthed gear.
You could use double isolation in the internal wires, make sure the circuit boards and other live parts have a distance of at least 10mm (preferably more) from the chassis, use extra isolation between the transformer and chassis, do use heat shrinks on all wires/connections, use heatsinks on the rectifier diodes instead of bolting them to the chassis and omit the ground-to-chassis connection. All at your own risk of course, but that also accounts for building an amplifier with a connection to earth.
A better (read safer) option would be to replace your mains wiring and add a earthed line to it. That's what I did a few years ago.
You could use double isolation in the internal wires, make sure the circuit boards and other live parts have a distance of at least 10mm (preferably more) from the chassis, use extra isolation between the transformer and chassis, do use heat shrinks on all wires/connections, use heatsinks on the rectifier diodes instead of bolting them to the chassis and omit the ground-to-chassis connection. All at your own risk of course, but that also accounts for building an amplifier with a connection to earth.
A better (read safer) option would be to replace your mains wiring and add a earthed line to it. That's what I did a few years ago.
I have read the F5 manual a few times more and think I understand the circuit better now: R9 is the gate osc prevention for resistor Q1, R10 sets the input impedance, Q1 is the voltage amplifier with a gain of 10+ (probably more, not sure how much). R3 and R1 sets the gain together with Q1´s ´internal resistance´. TH1 and R15 prevents thermal runaway. Q5/R17/R19/R21 limits the current by overbias or output short. R13 is the gate osc prevention resistor for Q3, and Q3 is the current amplifier. R11 stabilizes the circuit, and R5/R7 are FB resistors together with R3 (which I dont understand).
Understanding the circuit is part of the way. So does Q5 also securely limit that the output FETs does not get more than 20V input, which I have read in Bob Cordels book will kill it ?
In all the many posts that I have read, debugging has been a very minor subject. It is almost like: Build the thing and Hope and Pray
My idea was to test the circuit backwards: Desolder the Drain and Gate on the N-channel Power MOS FET and test if it works according to Nelsons document ´How to - Matching Devices, with an external PS. Then desolder the Source and Gate on the P-channel and test that too. Test the voltage across R11/R12 and test if it possible to get around 4V across R3 and R4. I dont know how to test the Q5 and Q6 current limiting circuits neither the thermistor circuits.
You could also use a signal from a signalgenerator or if such is not present, send an audio signal to the input to see if it gets amplified by Q1 / Q2. A standard DMM will probably not measure higher than between 100 - 1000 hz on AC mV, but that doesnt matter. You will still get a good indication. A scope would be better.
I have always been quite uncomfortable with solid state power amps, because I have had my accicents with some, and therefore I would prefer to test the board without the Power MOS FETs connected. Is that possible, or does the circuit need the Power devices to work correctly ? That would have been a good extension to the illustrated build manual.
As stated earlier I am mostly used to tube amps, and I would never install any tubes, before I have checked the voltages on all pins. And I allways test the PSU unloaded. That has saved me a lot of trouble over time.
... But I disconnected the PSU´s and it showed that the fancy pancy double diodes had died on me. I changed the contruction to Vishay 35A bridges - the PSU worked again, unloaded.
I then used the recommandation to use som smaller external PSU´s to test the circuits. It seems that both side work. No problem getting voltage across R3/4 and R11/12. I went up to drawing 0.5A in each side.
Quite happy I connected the newly working PSU in one channel - blowing the 1A fuse I had put in for safety. Ok I thought I will put in a lager one - 1,6A - also blown.
My F5 is allmost standard, but I have used the Toshiba MOS FETs and 2,2 mH coils in the power supply. Do the coils draw so much current that they will smelt the fuse? I did check the wiring, there does not seem to be anything wrong with that.
I am not that happy with just putting larger fuses in, without knowing what I am doing.
Understanding the circuit is part of the way. So does Q5 also securely limit that the output FETs does not get more than 20V input, which I have read in Bob Cordels book will kill it ?
In all the many posts that I have read, debugging has been a very minor subject. It is almost like: Build the thing and Hope and Pray
My idea was to test the circuit backwards: Desolder the Drain and Gate on the N-channel Power MOS FET and test if it works according to Nelsons document ´How to - Matching Devices, with an external PS. Then desolder the Source and Gate on the P-channel and test that too. Test the voltage across R11/R12 and test if it possible to get around 4V across R3 and R4. I dont know how to test the Q5 and Q6 current limiting circuits neither the thermistor circuits.
You could also use a signal from a signalgenerator or if such is not present, send an audio signal to the input to see if it gets amplified by Q1 / Q2. A standard DMM will probably not measure higher than between 100 - 1000 hz on AC mV, but that doesnt matter. You will still get a good indication. A scope would be better.
I have always been quite uncomfortable with solid state power amps, because I have had my accicents with some, and therefore I would prefer to test the board without the Power MOS FETs connected. Is that possible, or does the circuit need the Power devices to work correctly ? That would have been a good extension to the illustrated build manual.
As stated earlier I am mostly used to tube amps, and I would never install any tubes, before I have checked the voltages on all pins. And I allways test the PSU unloaded. That has saved me a lot of trouble over time.
... But I disconnected the PSU´s and it showed that the fancy pancy double diodes had died on me. I changed the contruction to Vishay 35A bridges - the PSU worked again, unloaded.
I then used the recommandation to use som smaller external PSU´s to test the circuits. It seems that both side work. No problem getting voltage across R3/4 and R11/12. I went up to drawing 0.5A in each side.
Quite happy I connected the newly working PSU in one channel - blowing the 1A fuse I had put in for safety. Ok I thought I will put in a lager one - 1,6A - also blown.
My F5 is allmost standard, but I have used the Toshiba MOS FETs and 2,2 mH coils in the power supply. Do the coils draw so much current that they will smelt the fuse? I did check the wiring, there does not seem to be anything wrong with that.
I am not that happy with just putting larger fuses in, without knowing what I am doing.
Just ordered a vario transformer, because I dont know anybody who has one I can borrow.
This project has been a quite long and expensive journey for me. I ordered the F5 kits in May 2011, and I am still struggeling to make it work
Probably the VDR i mentioned is called a PTC But it was delivered from Amplimo the transformer provider.
This project has been a quite long and expensive journey for me. I ordered the F5 kits in May 2011, and I am still struggeling to make it work
Probably the VDR i mentioned is called a PTC
But it was delivered from Amplimo the transformer provider.
I would suggest using the recommended fuse to start - 2.5A slow blow. And be sure to have the CL-60 thermistors in the primary circuit to suppress inrush.
If you use too low a value they will just blow. You are currently experiencing this.
From your most recent description it seems that you should have a working amp - what is the most vexing problem you have?
Also do you have a light bulb test lead? That will save the circuit from the most amazing errors.
Blowing fuses indicates you are not using your Mains Bulb Tester.
Blowing fuses probably indicates an assembly fault. That could be in wiring or a reversed component or a bias adjuster set in the wrong direction.
A Bulb Tester allows you to start up with some and all of these "faults" and importantly allows you to measure various points to find the anomaly.
Blowing fuses probably indicates an assembly fault. That could be in wiring or a reversed component or a bias adjuster set in the wrong direction.
A Bulb Tester allows you to start up with some and all of these "faults" and importantly allows you to measure various points to find the anomaly.