I had a problem in one channel , I compared components with the good channel and them desoldered the wrong one. This worked except for the output bjt.
Well done Antonio your enclosure is super impressive. Really unique and quite functional. I'm glad your happy with how it sounds. Please give us all a update once you have had time to listen to some of your favourite music.
Yes, bias at the start it goes up to 48-50mV then as the amp is getting hotter bias goes back to 43-44mV and then is stable.
I bought a Hakko FR-301 about 2 years ago and it has been a awesome. I had issues with applying heat and then quickly putting the sucker in position. This thing works wonders. Once in while I'll need to add solder before removing it, but very efficient. Well worth the expense.
I think all....but for me the most important about distortion it is the topology schematic....OS
So the Wolverine its a Great Amp for sure ....its a Neutral amp not a Musical amp....
Last edited:
Ok, ready for more dumb questions? I am so close to turning my Mouser project into a shopping cart, but....now that I have all the part numbers and quantities in there I started to think about matching transistors for performance. At Section 14 of the Build Guide (dated May 22, 2024 2nd Group Buy, but I have the 3rd Group Buy boards which I believe are identical) it says to match Q1 and Q2, and Q3 and Q4. But shouldn't Q5 and Q6 also be matched as all three pairs are bonded for thermal tracking?
With help from mainframe99 I finally decided on Onsemi NJW3201/NJW1302 output transistors. Should these be matched too? I don't see anything in the Build Guide on this.
I'll be using a Peak Atlas DCA Pro 75 to do the matching. How many of each type of transistor would you recommend I buy in order to get the Hfe/Vbe matches? I can't pull the trigger on my order until I know.
Do any other transistors need to be matched?
Separate question: I now realize that because I am mounting my EF3-4 boards parallel to the heatsinks, I can completely eliminate the 3mm thick heatsink with Q107 and Q108 mounted on it. I'll mount them on the main heatsink under the board and put Q103 right between them. I have the diyaudio 5U chassis with the standard UMS heatsinks. I haven't checked yet, but I am surmising that I may have to drill a few M3 holes to accommodate these transistors. Am I thinking about this correctly?
Many thanks,
John
PS: Did I mention that I get shaky when I'm not soldering😀.
With help from mainframe99 I finally decided on Onsemi NJW3201/NJW1302 output transistors. Should these be matched too? I don't see anything in the Build Guide on this.
I'll be using a Peak Atlas DCA Pro 75 to do the matching. How many of each type of transistor would you recommend I buy in order to get the Hfe/Vbe matches? I can't pull the trigger on my order until I know.
Do any other transistors need to be matched?
Separate question: I now realize that because I am mounting my EF3-4 boards parallel to the heatsinks, I can completely eliminate the 3mm thick heatsink with Q107 and Q108 mounted on it. I'll mount them on the main heatsink under the board and put Q103 right between them. I have the diyaudio 5U chassis with the standard UMS heatsinks. I haven't checked yet, but I am surmising that I may have to drill a few M3 holes to accommodate these transistors. Am I thinking about this correctly?
Many thanks,
John
PS: Did I mention that I get shaky when I'm not soldering😀.
See here on matching : Post 3476
Hope you didn't buy loads of these. Unlike sankens and toshibas, these will be crazy close 1-2% same type and about 5% between types, there is no need to worry.
This is good. Any builder can do wolverine with just a multimeter.
for Q1 2 3 4 5 6 I would suggest 50 min each would likely get you all the matches you need.
As linked above. Up to you. Only Q1 2 3 4 critical. I've done amps matching only these, and the amps amazing.
Yes, as I mentioned previously, you will need to drill and tap Q103, 107 and 108 holes 🙂
To match transistors, they need to be at exactly the same temperature. I mean exactly.
I gave the forum members a jig design to match transistors ages ago, several boards have been designed for it so far. "Testers" never get the matches close. The jig places two transistors in a diff pair configuration, you mount them together, they are run on a current source for tail current. You simply look for a null between the collectors (0.1% resistors). It makes a difference.
Touch a transistor or breathe on it, and your peak (or any meter type) matcher will not give you the same results, plus the jig is very sensitive. The only way you can match them closely is if they are run a the same time, same current and thermally isolated together. You put foam over the pair, hen a box over that to block air currents. Yup, transistors are that temperature sensitive.
I gave the forum members a jig design to match transistors ages ago, several boards have been designed for it so far. "Testers" never get the matches close. The jig places two transistors in a diff pair configuration, you mount them together, they are run on a current source for tail current. You simply look for a null between the collectors (0.1% resistors). It makes a difference.
Touch a transistor or breathe on it, and your peak (or any meter type) matcher will not give you the same results, plus the jig is very sensitive. The only way you can match them closely is if they are run a the same time, same current and thermally isolated together. You put foam over the pair, hen a box over that to block air currents. Yup, transistors are that temperature sensitive.
In certain locations, it is critical. It reduces distortion and DC drift.
You can lower distortion by matching complimentary drivers and other parts. Output stages with matched transistors exhibit lower distortion as well. If you have other problems you won't see improvement. If the circuit and build are performing well enough, you will. That assumes you have equipment good enough to measure down to those levels.
The diff pair is the error comparator. It can't do that well if it isn't very closely matched. Errors in following stages need to be corrected via feedback. The less correction needed, the lower the distortion and the better it sounds. It's terribly logical.
You can lower distortion by matching complimentary drivers and other parts. Output stages with matched transistors exhibit lower distortion as well. If you have other problems you won't see improvement. If the circuit and build are performing well enough, you will. That assumes you have equipment good enough to measure down to those levels.
The diff pair is the error comparator. It can't do that well if it isn't very closely matched. Errors in following stages need to be corrected via feedback. The less correction needed, the lower the distortion and the better it sounds. It's terribly logical.
I hope you see the same values 48-50 again in mv and them 44 mv steady and the winter. I probably have a problem with my amplifier. And I say this because it was regulated in the winter at 44 mv and now in the summer when I put him back on the bench I counted the maximum at 30mv bias with the same temp on heatsinks at 40C. at the two channels
...
I have 5 of the transistor matching boards available that are based on @anatech's design. $13 shipped in the lower 48 States of USA. PM me if interested. I can ship outside the lower 48 but it will cost more. I have used the board to match transistors on a numner of projects and it works EXTREMELY well.
Hi Guy's,
While I agree that transistor matching is important and worthwhile, I’d like to share a different perspective based on our experience during the prototype phase of the Wolverine development. At that time, we conducted distortion testing without performing any transistor matching at allg. I just looked back at the measurements and we were able to achieve a distortion level of 0.000285% at 1Khz with a bandwidth of 20kHz and 96kHz at 80W with an 8-ohm load.
So, my point is that while transistor matching is beneficial, it's not something to stress over if you can't achieve perfect matching.
Additionally, in simulation, it appears that matching transistors Q3-Q4 is the most critical, followed by Q1-Q2.
While I agree that transistor matching is important and worthwhile, I’d like to share a different perspective based on our experience during the prototype phase of the Wolverine development. At that time, we conducted distortion testing without performing any transistor matching at allg. I just looked back at the measurements and we were able to achieve a distortion level of 0.000285% at 1Khz with a bandwidth of 20kHz and 96kHz at 80W with an 8-ohm load.
So, my point is that while transistor matching is beneficial, it's not something to stress over if you can't achieve perfect matching.
Additionally, in simulation, it appears that matching transistors Q3-Q4 is the most critical, followed by Q1-Q2.
Hi Stuart,
Yes, if the design has intrinsic low distortion in the output and Vas stages, it will have low distortion for sure. As you also noted, matching closely can matter. Maybe it also depends on how much performance is important to the equipment. That low distortion makes no sense in a disco or rock concert. It may matter in a very high quality reproduction system. If your main speakers are Bose or Cerwin Vega then who cares for example? I'm sure I just offended some folks.
Yes, if the design has intrinsic low distortion in the output and Vas stages, it will have low distortion for sure. As you also noted, matching closely can matter. Maybe it also depends on how much performance is important to the equipment. That low distortion makes no sense in a disco or rock concert. It may matter in a very high quality reproduction system. If your main speakers are Bose or Cerwin Vega then who cares for example? I'm sure I just offended some folks.
Yeah there’s needs to be a line in the sand by the builder for transistor matching nervosa and just getting on with it and enjoying music. Some people enjoy the thrill of the chase to match. No problem.