Now all I have left to order is 2 transformers and 2-4 heatsinks ! 
I matched and sorted them. Almost all sets are within .01V at .48A, but a few are within .02V. The margins are likely the same because my meter (Triplett) only reads to .01V.
Here's my procedure:
I used a 12V linear regulated supply that I let warm up for 5 hrs, and I checked it often, and it stayed at 11.98V throughout the matching. The G and D were connected through 3 5ohm 50W resistors to the supply to the S. The DVM was across the fet to read Vgs. The fet was under current for 3 seconds to ensure that the DVM had stabilized, but a consistent time so the dissapation didn't heat the fet (more than all of the others) and drop the Vgs due to neg temp coef. In reality, the current was around .48A, but varied a little with Vgs. But the absolute Vgs isn't critical, only the Vgs variation between fets, and if their Vgs is the same, then their test current was the same.
All of the fets were numbered and tested. All of the raw data was entered in a spreadsheet and sorted. Then came the fun part, making sets. The distibution was a decent representation of a bell curve. Some people needed sets of 12 and some sets of 4, and some in between. Now that they are all sorted, I'll retest the sets to be sure that no mistakes were made before I ship.
I started to match them ar .70A ansl so I could calculate the tranconductance, but they were all too close to sort with any confidence. The change in Vgs was .13V or .14V for all, so I quit half way through. I planned to have a printout of the raw data for every numbered fet, but the floppy in my old laptop that I was using crapped out, so I can't intall a printer or save to a floppy and move to my main machine, so I'll sent hand written data.
I think everyone will be pleased. with 309 fets to choose from, There were a lot of combinations to choose from. I think you would be hard pressed to get tighter matches . I was literally pushing the limits of resolution on my $100 DVM.
I'll send a few scraps that fell outside of spec with each order. If you're making a larger aleph with lots of parellel fets, may I suggest something. first power it up and make sure the front end board is fully functional with one fet and source resistor per rail. It will play and let you make sure everything is working before you risk a matched set. (Don't ask why I know this and suggest this method)
If you haven't sent payment or at least contacted me, There are a few people who want your fets.
Here's my procedure:
I used a 12V linear regulated supply that I let warm up for 5 hrs, and I checked it often, and it stayed at 11.98V throughout the matching. The G and D were connected through 3 5ohm 50W resistors to the supply to the S. The DVM was across the fet to read Vgs. The fet was under current for 3 seconds to ensure that the DVM had stabilized, but a consistent time so the dissapation didn't heat the fet (more than all of the others) and drop the Vgs due to neg temp coef. In reality, the current was around .48A, but varied a little with Vgs. But the absolute Vgs isn't critical, only the Vgs variation between fets, and if their Vgs is the same, then their test current was the same.
All of the fets were numbered and tested. All of the raw data was entered in a spreadsheet and sorted. Then came the fun part, making sets. The distibution was a decent representation of a bell curve. Some people needed sets of 12 and some sets of 4, and some in between. Now that they are all sorted, I'll retest the sets to be sure that no mistakes were made before I ship.
I started to match them ar .70A ansl so I could calculate the tranconductance, but they were all too close to sort with any confidence. The change in Vgs was .13V or .14V for all, so I quit half way through. I planned to have a printout of the raw data for every numbered fet, but the floppy in my old laptop that I was using crapped out, so I can't intall a printer or save to a floppy and move to my main machine, so I'll sent hand written data.
I think everyone will be pleased. with 309 fets to choose from, There were a lot of combinations to choose from. I think you would be hard pressed to get tighter matches . I was literally pushing the limits of resolution on my $100 DVM.
I'll send a few scraps that fell outside of spec with each order. If you're making a larger aleph with lots of parellel fets, may I suggest something. first power it up and make sure the front end board is fully functional with one fet and source resistor per rail. It will play and let you make sure everything is working before you risk a matched set. (Don't ask why I know this and suggest this method)
If you haven't sent payment or at least contacted me, There are a few people who want your fets.
The number printed on the sticker is just the number of the fet. I gave each a number, measured Vgs, and entered the data into a spreadsheet. The only problem is I used my old 386 laptop. When I tried to save the data, I found thet the floppy had crapped, so I couldn't get the data out, or install a printer to print it. (as if I would have been able to find drivers for windows 3.1) So I used the laptop to sort, group. I had planned to print the raw data so everyone could see my methods.
Most sets are within .02 V. matched at .48A. Some will have a number like 4.12 hand written. This is thr Vgs as measured under my conditions. most will have a number 1 to 4. this is the group number. All groups were retested before packing to be sure I didn't blow one in the handleing or misrecord a number along the way. I also included extras for testing in every order. If you're building Alephs, I would recomend first testing the front end of the amp with 1 fet per rail and using a larger source resistor to be sure the amp is working propperly. It won't give anywhere near full current, but will give full voltage and show that it is all working before hooking up a full set of matched fets.
Enjoy
Most sets are within .02 V. matched at .48A. Some will have a number like 4.12 hand written. This is thr Vgs as measured under my conditions. most will have a number 1 to 4. this is the group number. All groups were retested before packing to be sure I didn't blow one in the handleing or misrecord a number along the way. I also included extras for testing in every order. If you're building Alephs, I would recomend first testing the front end of the amp with 1 fet per rail and using a larger source resistor to be sure the amp is working propperly. It won't give anywhere near full current, but will give full voltage and show that it is all working before hooking up a full set of matched fets.
Enjoy
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.