Amp is running
yes Hugh and I finally got the amp running after fixing the second issue. We also did some listening comparisons with one of Hugh's amps.
This should come as no surprise but there really was no comparison, Hugh really does have a magical setup I would recommend anyone serious about hifi to audition.
From what I have learnt of this layout to date are as follows
1) be careful as to the selection of caps, this layout is designed for specific brands. Check the lead pitch and diameter.
2) the C2 through holes should be 1.2mm to cater for silver mica leads. I had to drill these.
3) the resistor pads are only 1.4mm I think a little bigger would be better, I know we have through plating however there really is not much pad to play with.
4) Not a lot of room for the output inductor, I made mine with 1.6mm at 4.1 uH. And it really was a very tight squeeze with C12. Guess an off board zobel could be the solution.
5) Component designations on the layout would have been helpful
6) This is probably the most important; was the spacing between the o/p transistors is quite small and in particular the driver transistors were virtually hard up against the first o/p. I also found this amp getting quite hot in comparison to the original SC, whilst there are probably other contributing reasons for this I think the closeness of the o/p played a large part.
7) I am a little sceptical about the high current traces circling the board however I think this will show up in THD tests if it is an issue.
I have already started playing round with the layout to improve on these points however will not progress until I do the following.
I am intending on doing some THD measurements of the amps performance and since I have the original SC I will be able to draw comparisons.
I am also intending on measuring the thermal compensation of the thermalTraks internal diodes. Whilst the initial bios was easy to achieve with the mechanisms put in place the real test will be in its ability to track.
I think Carl has done a fantastic job in getting the amp to this stage for all of us to benefit, I for one really appreciate it.
A big thanks to Hugh for his help and wisdom, there really is no substitute for decades of experience, take the lid off one of his amps and you will see the workmanship and years of design that has gone into every part of it.
Peter.
yes Hugh and I finally got the amp running after fixing the second issue. We also did some listening comparisons with one of Hugh's amps.
This should come as no surprise but there really was no comparison, Hugh really does have a magical setup I would recommend anyone serious about hifi to audition.
From what I have learnt of this layout to date are as follows
1) be careful as to the selection of caps, this layout is designed for specific brands. Check the lead pitch and diameter.
2) the C2 through holes should be 1.2mm to cater for silver mica leads. I had to drill these.
3) the resistor pads are only 1.4mm I think a little bigger would be better, I know we have through plating however there really is not much pad to play with.
4) Not a lot of room for the output inductor, I made mine with 1.6mm at 4.1 uH. And it really was a very tight squeeze with C12. Guess an off board zobel could be the solution.
5) Component designations on the layout would have been helpful
6) This is probably the most important; was the spacing between the o/p transistors is quite small and in particular the driver transistors were virtually hard up against the first o/p. I also found this amp getting quite hot in comparison to the original SC, whilst there are probably other contributing reasons for this I think the closeness of the o/p played a large part.
7) I am a little sceptical about the high current traces circling the board however I think this will show up in THD tests if it is an issue.
I have already started playing round with the layout to improve on these points however will not progress until I do the following.
I am intending on doing some THD measurements of the amps performance and since I have the original SC I will be able to draw comparisons.
I am also intending on measuring the thermal compensation of the thermalTraks internal diodes. Whilst the initial bios was easy to achieve with the mechanisms put in place the real test will be in its ability to track.
I think Carl has done a fantastic job in getting the amp to this stage for all of us to benefit, I for one really appreciate it.
A big thanks to Hugh for his help and wisdom, there really is no substitute for decades of experience, take the lid off one of his amps and you will see the workmanship and years of design that has gone into every part of it.
Peter.
Peter,
I am glad that you have your amplifier working and appreciate your comments. You are running a third more output devices with this amplifier than what you had in the published SC design. I would think it would be reasonable to expect that you would need a third more heatsink area for similar thermal behaviour.
What voltage are you running on your power supply rails? Do you have a way of measuring power output?
I am especially interested in seeing how well this thermal compensation scheme works. Do you have any comments on that?
I am glad that you have your amplifier working and appreciate your comments. You are running a third more output devices with this amplifier than what you had in the published SC design. I would think it would be reasonable to expect that you would need a third more heatsink area for similar thermal behaviour.
What voltage are you running on your power supply rails? Do you have a way of measuring power output?
I am especially interested in seeing how well this thermal compensation scheme works. Do you have any comments on that?
Peter,
It was a privilege to meet and work with you. I like your logical, measured approach - the only way to do this sort of work. You can't see, hear or feel electrons, and much of what we do is deduction based on measurements. You have the intuition and clearly DIY fascinates you. Your build quality was outstanding, and the sound we had pretty good.
Carl,
You are right about dissipation; the heatsink needs to be a little bigger. However, there were no thermal issues, though the drivers are a little close to the outputs and this could cause assembly issues with insulating washers.
I did not have enough time to measure off the thermal track of the diodes. My impression is that with 0.22R resistors all is well. However, it would take a lot to beat a BD139 with 1.5K from collector to base and roughly 430R from base to emitter. Placed on top of one of the outputs, and bypassed by a quality 22uF cap, this is a very effective Vbe multiplier. One can also place a 22uF cap across the 1K5 resistor from C to B and improve its performance further, too.
This is a good, basic amplifier and with a competent power supply sounds pretty good. Difficult to improve upon for the money.
Cheers,
Hugh
It was a privilege to meet and work with you. I like your logical, measured approach - the only way to do this sort of work. You can't see, hear or feel electrons, and much of what we do is deduction based on measurements. You have the intuition and clearly DIY fascinates you. Your build quality was outstanding, and the sound we had pretty good.
Carl,
You are right about dissipation; the heatsink needs to be a little bigger. However, there were no thermal issues, though the drivers are a little close to the outputs and this could cause assembly issues with insulating washers.
I did not have enough time to measure off the thermal track of the diodes. My impression is that with 0.22R resistors all is well. However, it would take a lot to beat a BD139 with 1.5K from collector to base and roughly 430R from base to emitter. Placed on top of one of the outputs, and bypassed by a quality 22uF cap, this is a very effective Vbe multiplier. One can also place a 22uF cap across the 1K5 resistor from C to B and improve its performance further, too.
This is a good, basic amplifier and with a competent power supply sounds pretty good. Difficult to improve upon for the money.
Cheers,
Hugh
Hello parb,
Not much has happened. I have updated the gerbers to improve some of the parts clearances that were commented on. Altho I had built a previous version of this amplifier and had enjoyed the results, I have never built this version. Peter had offered to report his measurements and findings but as of this date I have not heard from him.
Not much has happened. I have updated the gerbers to improve some of the parts clearances that were commented on. Altho I had built a previous version of this amplifier and had enjoyed the results, I have never built this version. Peter had offered to report his measurements and findings but as of this date I have not heard from him.
I am currrently building one of these amps for my brother and will post results when finished. I have made a few subtle changes to the original.
.Replaced beta enhancer bc639 with 2sc2240.
.current mirror = 2x mpsa18
.27 ohms in the VAS emitter
.VAS transistor = 2sc3423, VAS currect source = 2sa1360
.replaced MJE drivers with Toshiba 2sa1930 / 2sc5171
.changed driver emitter resistors to a single 220 ohm as per D Self recommendation.
.changed LTP emitter resistors from 100 ohm to 150 ohm
.reduced Cdom from 100pf ceramic to 68pf silver mica
.changed current mirror emitter resistors to 100 ohm
.Added a 3rd transistor to the current mirror for LTP collector current and voltage matching.
.Added bias trim mod
.Replaced beta enhancer bc639 with 2sc2240.
.current mirror = 2x mpsa18
.27 ohms in the VAS emitter
.VAS transistor = 2sc3423, VAS currect source = 2sa1360
.replaced MJE drivers with Toshiba 2sa1930 / 2sc5171
.changed driver emitter resistors to a single 220 ohm as per D Self recommendation.
.changed LTP emitter resistors from 100 ohm to 150 ohm
.reduced Cdom from 100pf ceramic to 68pf silver mica
.changed current mirror emitter resistors to 100 ohm
.Added a 3rd transistor to the current mirror for LTP collector current and voltage matching.
.Added bias trim mod
I will post the schematic in a few days when I've finalised all mods. I'm using the original SC PCB so I've had to cut a few tracks and create a few new ones along the way.
Looking back at the earlier postings, most of my mods fall into the already accepted category. The 3 transistor mirror cct was discussed in the mirror thread and the device substitutions are all no brainers.
As was pointed out by aksa, the emitter resistors should be 0.22R each rather than 0.1R to avoid an overly high level of idle dissipation for optimum PA quiescent. However, I'm not sure what the effect (or extent thereof) will be on tempco or even if the original values of RE were optimum - so I might need to make adjustments - with series + parallel resistance to the diode string. I currently have a 220R Bourne trimmer mounted across a 10R resistor for 'tweaking' and I will eventually fit a single resistor(s).
Looking at the number of design corrections and upgrades from the original I'm looking forward to hearing the end result. Hopefully only a weekend or 2 before that happens.
If it's worthy then yes, that would be great. I'll keep you posted.
I have some very good speakers for evaluation, so I'll know pretty quickly how good it is.
Cheers, vm
Transitional miller comp??
I thought of something along those lines about 10 years ago - with a resistor from the output feeding a junction between 2 miller comp caps. The VAS was a high beta CFP type and I implemented it in a preamp. However, I struggled to hear any audible improvement - but then again it was a class A preamp. Perhaps it's worth revisiting within a class AB power amp? Those sims look good and I'm surprised that the 10kHz plot looks so good too. I wonder though if there are limitations when driving real world speakers. Those sims I gather were with a 8+j0 load??
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.