I checked and the Buffalo 1 is ok. I'll take one. Thanks...
1. cersepn x 1
2. icewall x 2
3. pmchoong x 1
4. qusp x 1 this is tentative mate, but wanted to make sure my name was down
5. regal
6. randytsuch x 1
7. LK x1
8. greyrab x1
9. alexcount x1
10. adelias x1
11. Coolhead x1
12. Emphrygian x 1
13. tranhieu x 1
14. crazyfrog x 1
15. woresseo x 2
16. Pedefede x 1
17. Algar_emi x 1
18. ggetzoff x 1
1. cersepn x 1
2. icewall x 2
3. pmchoong x 1
4. qusp x 1 this is tentative mate, but wanted to make sure my name was down
5. regal
6. randytsuch x 1
7. LK x1
8. greyrab x1
9. alexcount x1
10. adelias x1
11. Coolhead x1
12. Emphrygian x 1
13. tranhieu x 1
14. crazyfrog x 1
15. woresseo x 2
16. Pedefede x 1
17. Algar_emi x 1
18. ggetzoff x 1
Will it work with Buffalo III? Is there a place that I can buy all the components or is QUSP offering those parts in a package?
Thanks
1. cersepn x 1
2. icewall x 2
3. pmchoong x 1
4. qusp x 1 this is tentative mate, but wanted to make sure my name was down
5. regal
6. randytsuch x 1
7. LK x1
8. greyrab x1
9. alexcount x1
10. adelias x1
11. Coolhead x1
12. Emphrygian x 1
13. tranhieu x 1
14. crazyfrog x 1
15. woresseo x 2
16. Pedefede x 1
17. Algar_emi x 1
18. ggetzoff x 1
19. syklab x1 + fets
Thanks
1. cersepn x 1
2. icewall x 2
3. pmchoong x 1
4. qusp x 1 this is tentative mate, but wanted to make sure my name was down
5. regal
6. randytsuch x 1
7. LK x1
8. greyrab x1
9. alexcount x1
10. adelias x1
11. Coolhead x1
12. Emphrygian x 1
13. tranhieu x 1
14. crazyfrog x 1
15. woresseo x 2
16. Pedefede x 1
17. Algar_emi x 1
18. ggetzoff x 1
19. syklab x1 + fets
Last edited:
I'll make it 20...
1. cersepn x 1
2. icewall x 2
3. pmchoong x 1
4. qusp x 1 this is tentative mate, but wanted to make sure my name was down
5. regal
6. randytsuch x 1
7. LK x1
8. greyrab x1
9. alexcount x1
10. adelias x1
11. Coolhead x1
12. Emphrygian x 1
13. tranhieu x 1
14. crazyfrog x 1
15. woresseo x 2
16. Pedefede x 1
17. Algar_emi x 1
18. ggetzoff x 1
19. syklab x1 + fets
20. Magsy x 1
1. cersepn x 1
2. icewall x 2
3. pmchoong x 1
4. qusp x 1 this is tentative mate, but wanted to make sure my name was down
5. regal
6. randytsuch x 1
7. LK x1
8. greyrab x1
9. alexcount x1
10. adelias x1
11. Coolhead x1
12. Emphrygian x 1
13. tranhieu x 1
14. crazyfrog x 1
15. woresseo x 2
16. Pedefede x 1
17. Algar_emi x 1
18. ggetzoff x 1
19. syklab x1 + fets
20. Magsy x 1
i might be able to find some that are decent matches that are better than Nelsons guidelines still, but no more perfect matches and i wont be doing any more matching, at least thats how i feel now, last time was an epic effort that i do not wish to, or have time to do again i'm afraid.
the sink needs to be able to get rid of roughly 50W, while keeping case temp under ~65c
the sink needs to be able to get rid of roughly 50W, while keeping case temp under ~65c
Ok, so thats a thermal resistance of about 1 C/W .... But those values are measured under quite different positioning, than the pictures i've seen, with heatsink attached to the bottum, and fins pointing down.
Does anyone have the thermal resistance of the heatsinks you guys used?
how are you matching them? vds will always be close, but just a reminder they should be matched at the operating point and for transconductance. if you are just doing VDS then i already have plenty matched for VDS, the whole rails were pretty close for VDS.
the sets i sold were matched under the conditions below in a circuit that is basically one arm of the D1, much tighter conditions than what Nelson and some others recommend for power amps
the sets i sold were matched under the conditions below in a circuit that is basically one arm of the D1, much tighter conditions than what Nelson and some others recommend for power amps
Attachments
Last edited:
I'm using the transconductance method you posted ealier, and exact the same circuit, as shown on the top portion of your diagram. I'm using 3 DVM (4 and 5 digits resolution, very good Fluke meters), a signal generator and a lab grade power supply. +45V supply, FET mounted on an heatsink using a clamp, Film caps. For the DC measurements, I adjust VR1 to get 24V on the 200R(30W) R1 resistor (exactly Id of 120ma), Let stabilze the FET then I note the VR1 wiper DC voltage. Then I apply a 1Khz, sinusoidal ac voltage at the gate using a 10uF film coupling cap and adjust its amplitude to get 3Vac at R1. Then I measured the Vgs at this point. I calculate the gm using an excel file...I'll post a picture of my setup tonight.
I was surprised to find that the first 5 I tested were very close to each other. Variability is very low, much better than a 2SK170-BL for example.
I was surprised to find that the first 5 I tested were very close to each other. Variability is very low, much better than a 2SK170-BL for example.
nice work! your main enemy will be room temp. i had sections of the rails of fets that were like that too, but unfortunately at least for me, it didnt stay consistent throughout different rails of fets, or really the rails themselves. there is definitely sections and some of the best quads i got were right next to each other in the rails consecutively. do remember to balance the 3 factors, for example a quad may be bettered by 2 pairs that are higher gM, as left right channel matching, though nice to have we are not that sensitive to it, so if you can have a lower distortion overall by having 2 pairs. obviously perfect quads of high gM are ideal, but not that common. i tested to 5 decimals though, probably overkill and only reliable to 3.5-4
I recommend going back and pulling a sample or 2 to test as a yard post if you are testing at a different time
I recommend going back and pulling a sample or 2 to test as a yard post if you are testing at a different time
Last edited:
Hi Algar,
Wow! Great work on the test setup. Both qusp and myself can attest to how much work it is to get it up and running correctly.
It's actually a really great setup for generally testing mosfets, so if you have the option, it's nice to keep it around.
I measured a lot of fets to find the ones I did, but I'm sure there are probably others out there with even better transconductance figures at this operating point. If you have anything laying around, run it through the test setup for fun and see what you get.
Cheers,
Owen
Wow! Great work on the test setup. Both qusp and myself can attest to how much work it is to get it up and running correctly.
It's actually a really great setup for generally testing mosfets, so if you have the option, it's nice to keep it around.
I measured a lot of fets to find the ones I did, but I'm sure there are probably others out there with even better transconductance figures at this operating point. If you have anything laying around, run it through the test setup for fun and see what you get.
Cheers,
Owen
indeed i still have and use it for matching fets. i do need to lash up a more flexible power supply for it though and ideally i would like to case it up to help keep temperature fluctuations to a minimum, along with actively heating the sink and monitor with a thermocouple. i'll eventually spring for better tempco than caddock on R1
I would have killed for better DVM's and a lab supply, then might not have had to build the extensive rig i did to augment and amplify the measurements. wasnt prepared to buy too much new equipment for the purpose though and the initial setup i had was far too sensitive to DC leakage from the 10uf film cap.
it was a total PITA, but i'm actually thankful for the thought experiment
Last edited:
Thanks to you guys for sharing the design, test circuit, procedure, etc.
I couldn't do it without it
. I'll post my preliminary results soon, just to make sure that my readings are making sense, but everything seems to be stable and accurate.
I'm using the same heatsink that I'll use for the actual dac, but I'll probably end up cutting it two (one for me and one for my friend). For R1, I'm still waiting after the exact Caddock TO220, hence using a bunch of 8 big WW 25 ohms 30W resistors. They are barely warm after stabilisation. Since we want to measure the FET and not the rest of the circuit, the more stable the circuit, the better are the measurements I guest.
My final testing when matching FET is always to build tha actual circuit, install FET on socket and test them again, insuring that the DC and AC match is indeed correct then I'll choose the pair (if possible the quad) with the lowest THD.
I have an extensive test lab at home that I built over the years, and worked in calibration labs, testing grounds and radar for most of my life.
Testing part for me is just routine and with good test equipment it helps a lot I include a picture of the test bench.
I couldn't do it without it
. I'll post my preliminary results soon, just to make sure that my readings are making sense, but everything seems to be stable and accurate.I'm using the same heatsink that I'll use for the actual dac, but I'll probably end up cutting it two (one for me and one for my friend). For R1, I'm still waiting after the exact Caddock TO220, hence using a bunch of 8 big WW 25 ohms 30W resistors. They are barely warm after stabilisation. Since we want to measure the FET and not the rest of the circuit, the more stable the circuit, the better are the measurements I guest.
My final testing when matching FET is always to build tha actual circuit, install FET on socket and test them again, insuring that the DC and AC match is indeed correct then I'll choose the pair (if possible the quad) with the lowest THD.
I have an extensive test lab at home that I built over the years, and worked in calibration labs, testing grounds and radar for most of my life.
Testing part for me is just routine and with good test equipment it helps a lot
I include a picture of the test bench.Attachments
Working for 7 years in a calibration labs, that eventually was closed down (that was the bad part) allowed me at the time to buy used test gears they were getting rid of at ridiculous price. The good part was that I new they were all in great shape. So I lost my jobs but I was able to stock some nice gears. It was hard to buy at the time but worth it over the long run. I have four scopes for example
Here my FET test setup. Nothing fancy really, a breadboard, a few parts and the test bench.
Here my FET test setup. Nothing fancy really, a breadboard, a few parts and the test bench.
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.