@ugly_woofer said:
Did you get the DC offset down ?
Not yet. I'm working on that right now. DC offset at idle is running at about 45mv in both channels, which is a little on the high side. I'd like to get that down to less than 10mv if I can. One way to accomplish this would be to install Tony's adjustable DC offset mod with 5k 10 turn trimmer pots. The other way would be to tweak the differential pair current source and feedback resistors slightly to eliminate the offset.
I was able to get the DC offset down from 45mv to about 9mv by tweaking the differential pair current source and feedback resistors slightly. This tweaking also increased the amount of negative feedback and reduced the closed loop gain of the amplifier from about 40dB down to 35dB. Power output before clipping is now slightly higher at 10.5 watts rms at 1kHz and the outputs continue to run nice and cool with the new heatsinks. With the rails at +/-18vdc, the crossover notch distortion is also completely gone (due to the 470k resistor mod). I did a frequency response sweep with my scope and function generator and the amp seems fairly flat with just a small amount of rolloff at the extremes of 20Hz and 20Khz. So I think I am ready to button this project up by installing the transformer and doing the final wiring of all the pots and jacks.
I figured that as long as I was spending roughly 100 bucks on this project, I might as well go ahead and do xraytonyb's power supply mods as well. The kit came with small 1 amp 1N4007 diodes and 1000uF 35V filter caps. I replaced them with 3amp, 400piv power switching diodes (Mouser 625-RGP30G-E3/73) and Panasonic 2200uF 35V 105C caps (Mouser 667-EEU-FRV222LPF). The diodes were a little bigger, so I stagger mounted them above the board for cooling. The four new caps were the same diameter, but a little taller than the originals, so I etched the copper from a small piece of PCB material and silicone glued it to the top of the four caps. This provides a stable mounting for the four taller caps, yet is easy to remove if I need to for servicing.
@Kornbread said:
Of course, listen for a time and let us know how it sounds.
Will do. I have been playing music between some of the mods, but when I do I have to disconnect my function generator hookup and re-attach a 3.5mm plug to do so, which is kind of a hassle. I think what I need to do is build a little switching box so I can quickly switch back and forth between function generator and phone input. So far, it has been sounding really good when I play music from my phone. The gain at the input is quite high, so it picks up some hum from the relatively long, unshielded input cable that I am using. But this should improve considerably when I shorten this cable and connect it to the volume control pots.
I ran some detailed frequency response measurements with my scope and discovered that the low frequencies were rolling off a little faster than I would like. The -3dB points came out at 14Hz on the low end and 49kHz at the upper end. 20Hz was down 1.7dB while 20kHz was only down by about 0.5dB. I found that by boosting C2 and C5 from 47uF up to 100uF, this improved the low frequency extension considerably. 20Hz is now only down by about 0.5dB.
I need to get REW set up with my dummy load and my sound card using L-PAD type protection cables, but that is still a work in progress. Then I will be able to look at the harmonic distortion profile (2nd, 3rd, 4th, etc.) at various power levels.
This afternoon, I greased up and attached the aluminum angle bracket to the main bud box. After gobbing on a thick layer of heatsink grease, I used Eric's suggested razor blade squeegee method to thin it down to fill the imperfections only. Easier said than done, as I think I needed a wider razor blade to do the best job. I think it turned out OK, though (see pics). There was very little squeeze out when I bolted it together, so cleanup was easy-peasy.
This project is now complete. I ran it through a full set of tests and everything looks good. Using the Triad 24VCT 3.3a transformer, the rails idle at +/- 18.2vdc with an output bias current of 3.6ma. Ramping it up to full power before clipping at 1kHz produces 8.26Vrms in each channel, driving into an 8 ohm dummy load (or about 8.5watts rms/ch). At 20Hz, power output before clipping drops to 7.22Vrms (or about 6.5 watts rms). Not exactly a super power amp by any means, but this is roughly what I expected when I started this project.
Hum and noise: I boosted the 4 main filter capacitors from 1000uF to 2200uF, but the ac ripple is still fairly high. There is about 2.7vac of 120Hz ripple riding on the +/- 18.2vdc supply rails. This produces a constant hum that is clearly audible between music tracks. But it is not too bad. It does not change in amplitude or go away when I short the inputs, so I know this hum is coming from power supply ripple and not being induced by poor input wire dressing.
Heatsinks: I have run the amp at high power levels for an extended period and my aluminum angle + case heatsink arrangement seems to be working very well. Spot checks show the heatsinks and transistor cases to be running from 90F to 116F. There is about a 4 to 6F degree temperature differential between the top of each transistor and the nearby heatsink, which indicates that I am getting a good thermal transfer from case to sink. In addition, the transformer only gets warm to the touch and measures between 98F to 100F after running for a while.
Would I recommend the purchase of this kit? After strugging to get it working, I would have to give this kit a big thumbs down. One of the main problems was the poor quality circuit board. I have been running my soldering iron at about 650F, which is not that high, and I have encounted at least 6 places where the excessively small solder pads either broke or twisted off. I've got "iffy" looking patches rigged up all over the place. In addition, I had to replace my set of output transistors due to overheating problems.
On the positive side, this was a good training experience for me and I learned quite a few new things about how amplifiers work (or do not work). Thanks to xraytonyb for his excellent training video series. And thanks again to everyone here for all the positive feedback and suggestions along the way.
Bill
I made a small circuit board for all the input pot connections. The board is held in place by the six PCB mounting tabs on the pots.
I'm glad you got yours running reasonably well. I certainly agree with you giving this kit the thumbs down. The poor quality board would be a nightmare for someone new to electronics. My Hakko desoldering gun sucked a few pads off when I replaced the small transistors. Seems the power supply is not well designed and the counterfeit TO-3 transistors are the final nail in the coffin. I did learn some things from Tony's video series. I'm glad the kit was inexpensive. I won't feel so bad if I toss it in the circular file.
Which Hakko do you have? I have the FR-301 and I was afraid to go anywhere near this circuit board with it, even at the lowest heat setting. I burnt and lifted several pads with a very small tip set at 650F. And I twisted two pads off the board when tightening the screws on those two blue PCB mount screw terminal block connectors.
I have the 301 too. I leave mine set at 2 most of the time, but the temp range listed in the manual is really broad. It could have been as high as 700F.
Comments
Got it. As you describe it, I think this might work. Thanks, Steve!
Did you get the DC offset down ?
Not yet. I'm working on that right now. DC offset at idle is running at about 45mv in both channels, which is a little on the high side. I'd like to get that down to less than 10mv if I can. One way to accomplish this would be to install Tony's adjustable DC offset mod with 5k 10 turn trimmer pots. The other way would be to tweak the differential pair current source and feedback resistors slightly to eliminate the offset.
I was able to get the DC offset down from 45mv to about 9mv by tweaking the differential pair current source and feedback resistors slightly. This tweaking also increased the amount of negative feedback and reduced the closed loop gain of the amplifier from about 40dB down to 35dB. Power output before clipping is now slightly higher at 10.5 watts rms at 1kHz and the outputs continue to run nice and cool with the new heatsinks. With the rails at +/-18vdc, the crossover notch distortion is also completely gone (due to the 470k resistor mod). I did a frequency response sweep with my scope and function generator and the amp seems fairly flat with just a small amount of rolloff at the extremes of 20Hz and 20Khz. So I think I am ready to button this project up by installing the transformer and doing the final wiring of all the pots and jacks.
Of course, listen for a time and let us know how it sounds.
I figured that as long as I was spending roughly 100 bucks on this project, I might as well go ahead and do xraytonyb's power supply mods as well. The kit came with small 1 amp 1N4007 diodes and 1000uF 35V filter caps. I replaced them with 3amp, 400piv power switching diodes (Mouser 625-RGP30G-E3/73) and Panasonic 2200uF 35V 105C caps (Mouser 667-EEU-FRV222LPF). The diodes were a little bigger, so I stagger mounted them above the board for cooling. The four new caps were the same diameter, but a little taller than the originals, so I etched the copper from a small piece of PCB material and silicone glued it to the top of the four caps. This provides a stable mounting for the four taller caps, yet is easy to remove if I need to for servicing.
Will do. I have been playing music between some of the mods, but when I do I have to disconnect my function generator hookup and re-attach a 3.5mm plug to do so, which is kind of a hassle. I think what I need to do is build a little switching box so I can quickly switch back and forth between function generator and phone input. So far, it has been sounding really good when I play music from my phone. The gain at the input is quite high, so it picks up some hum from the relatively long, unshielded input cable that I am using. But this should improve considerably when I shorten this cable and connect it to the volume control pots.
In for a pound . . . (err - 2.8 lbs to be exact)
I ran some detailed frequency response measurements with my scope and discovered that the low frequencies were rolling off a little faster than I would like. The -3dB points came out at 14Hz on the low end and 49kHz at the upper end. 20Hz was down 1.7dB while 20kHz was only down by about 0.5dB. I found that by boosting C2 and C5 from 47uF up to 100uF, this improved the low frequency extension considerably. 20Hz is now only down by about 0.5dB.
I need to get REW set up with my dummy load and my sound card using L-PAD type protection cables, but that is still a work in progress. Then I will be able to look at the harmonic distortion profile (2nd, 3rd, 4th, etc.) at various power levels.
This afternoon, I greased up and attached the aluminum angle bracket to the main bud box. After gobbing on a thick layer of heatsink grease, I used Eric's suggested razor blade squeegee method to thin it down to fill the imperfections only. Easier said than done, as I think I needed a wider razor blade to do the best job. I think it turned out OK, though (see pics). There was very little squeeze out when I bolted it together, so cleanup was easy-peasy.
This project is now complete. I ran it through a full set of tests and everything looks good. Using the Triad 24VCT 3.3a transformer, the rails idle at +/- 18.2vdc with an output bias current of 3.6ma. Ramping it up to full power before clipping at 1kHz produces 8.26Vrms in each channel, driving into an 8 ohm dummy load (or about 8.5watts rms/ch). At 20Hz, power output before clipping drops to 7.22Vrms (or about 6.5 watts rms). Not exactly a super power amp by any means, but this is roughly what I expected when I started this project.
Hum and noise: I boosted the 4 main filter capacitors from 1000uF to 2200uF, but the ac ripple is still fairly high. There is about 2.7vac of 120Hz ripple riding on the +/- 18.2vdc supply rails. This produces a constant hum that is clearly audible between music tracks. But it is not too bad. It does not change in amplitude or go away when I short the inputs, so I know this hum is coming from power supply ripple and not being induced by poor input wire dressing.
Heatsinks: I have run the amp at high power levels for an extended period and my aluminum angle + case heatsink arrangement seems to be working very well. Spot checks show the heatsinks and transistor cases to be running from 90F to 116F. There is about a 4 to 6F degree temperature differential between the top of each transistor and the nearby heatsink, which indicates that I am getting a good thermal transfer from case to sink. In addition, the transformer only gets warm to the touch and measures between 98F to 100F after running for a while.
Would I recommend the purchase of this kit? After strugging to get it working, I would have to give this kit a big thumbs down. One of the main problems was the poor quality circuit board. I have been running my soldering iron at about 650F, which is not that high, and I have encounted at least 6 places where the excessively small solder pads either broke or twisted off. I've got "iffy" looking patches rigged up all over the place. In addition, I had to replace my set of output transistors due to overheating problems.
On the positive side, this was a good training experience for me and I learned quite a few new things about how amplifiers work (or do not work). Thanks to xraytonyb for his excellent training video series. And thanks again to everyone here for all the positive feedback and suggestions along the way.
Bill
I made a small circuit board for all the input pot connections. The board is held in place by the six PCB mounting tabs on the pots.
Transformer and power inlet connections:
The finished amplifier with top cover:
Good looking work, 4thtry - impressive execution.
Living up to your name's sake, I see.
I'm glad you got yours running reasonably well. I certainly agree with you giving this kit the thumbs down. The poor quality board would be a nightmare for someone new to electronics. My Hakko desoldering gun sucked a few pads off when I replaced the small transistors. Seems the power supply is not well designed and the counterfeit TO-3 transistors are the final nail in the coffin. I did learn some things from Tony's video series. I'm glad the kit was inexpensive. I won't feel so bad if I toss it in the circular file.
Which Hakko do you have? I have the FR-301 and I was afraid to go anywhere near this circuit board with it, even at the lowest heat setting. I burnt and lifted several pads with a very small tip set at 650F. And I twisted two pads off the board when tightening the screws on those two blue PCB mount screw terminal block connectors.
I have the 301 too. I leave mine set at 2 most of the time, but the temp range listed in the manual is really broad. It could have been as high as 700F.