• These commercial threads are for private transactions. diyAudio.com provides these forums for the convenience of our members, but makes no warranty nor assumes any responsibility. We do not vet any members, use of this facility is at your own risk. Customers can post any issues in those threads as long as it is done in a civil manner. All diyAudio rules about conduct apply and will be enforced.

GB 2SK3497/2SJ618 for EUVL F5X or other projects

Status
Not open for further replies.
Jump to Latest
#141
I normally would not have made any noise about the contact resistance.

But we have 2 extreme situations here :

1) The matching current is very high.
2) These mosfets have a huge transcoductance.

The combined error is therefore not negligible.

If it were for the amp I would have said put them in the oven (with wife's permission) and match at 50°C.
But for regulators the requirements are not so stringent.


Happy matching,
Patrick
 
#145
NicMac said:
Great work!
You are really doing all the work on the PRP GB. I will have to find a way to compensate......
Anyway, maybe it would be best divide all into groups of 2 channels.
Happy matching,
Nic
Click to expand...
No worry, Nic. The bigger the pot, the better match.

After I finished the 220R, I implemented a small program to optimize the distribution 😀 For example, Jan will get 4 sets of group of 20 (5 channels). Greg will get 4 sets of 16 (4 channels). You and me will get 4 sets of group 12 (3 channels). The rest will get 4 sets of 8 (2 channels). I hope the 33R will have similar yield, that will be excellent.
 
#147
Last night, I sampled 10 pairs and found the N-P didn't match.
It bothered me through out the night 🙁 This morning, I found Patrick's post

http://www.diyaudio.com/forums/pass-labs/183362-f5x-euvl-approach-27.html#post2632016

and I was happy again and in the mood to play more 😉

With no insulation between the mosfet and the heatsink block, the mosfet was barely warm. But then I thought the mosfets are not going to operate in such luxury in the real life, so I added a kapton film in between which promptly raised the temp to approx 40C. Obviously, as the temp goes up, the VGS was reduced.

One question to the expert, should I match the mosfets with better heat management or an artificial slightly crippled heat management? I would think the later because it is closer to the real operating condition. Any comments?

Thanks!!
 
#148
Fred,

You are going to use these MOSFETs not in the amp but in the regulator. There, the rail voltage equals (Zener diode voltage - Vgs of MOSFET at 4A). For a power amp with +/-16V rails, 0.2~0.4V volt difference is quite OK (2.5%). And the zener diodes are also not specified to better than a % or two.

So I would just match at room temperature and real quick (so no chance to heat up). The use of a big aluminium block is to keep the temperature at room temperture even when passing some 10W to the mosfets. You want as good thermal contact to the AL block as possible, but no need to insulate (i.e. connected to Drain) as long as the AL block is insulated from Gnd and whatever else.

As already mentioned, Vgs difference between N & P can be at least partially compensated by hand selecting Zener diodes. Just place them on sockets first so that you can swap them in and out.

For those who really want the rail voltages to be exactly 16.00V, they can replace the Zener with a trimmer, and trim in circuit (and inside the case) once everything is working, and at steady state.

But I do not consider it necessarily. The request to match in this particular case is to make sure you do not have too much variation to make up later. One can even argue it is not necessary.


Patrick
 
#149
Thank you, Patrick.

Since I will only do one pass of measurement, I will keep the MOSFET on the heatsink until the VGS is totally settled for at least a few seconds. With this heatsink, there is no chance to over heat the FET. They are in safe hands 🙂
 
#150
One small update, all the mosfets are matched and bagged. I went for best matches rather than high yield. I am pleased to announce that all the N-N, P-P are matched to <= 2 mv Vgs at 4A. I still have an extra 3-4 sets but are in the 3-5 mv range.

I also received shipment notification from MH&W today that the T-220 pads are in UPS's hand. If nothing else, I will package up the Christmas gift bags and ask for $$ very soon 🙂

The spreadsheet is reopen. I added 2 columns:
1. The destination country (this will help me to calculate the postage).
2. Shipment preference - ship directly to you, or ship with other GB items.
Please fill in the info.

Also, I will have a few matched 2SK2013/2SJ313 for BA3 and Juma TF5 project.
If any of you are interested in that, let me know. First come first serve as I don't know how many triplets I will have.

Thanks!!
 
#152
Thank you very much, Fred,

I'd be interested in four triplets of 2SK2013/2SJ313.

A note on the spreadsheet:

There will be two distributors for F5X related stuff in Europe:

smyslow: cases, heatsink, Furutech IEC

me (heinz1): Fukushima and PRP resistors

You can decide now whether your goods bought from Fred are shipped together with the resistors. Shipping charges for mailing within Europe were already collected from Horio during the Fuku GB; I can send up to 500g of goods for the charge you payed already.
 
#153
Thanks for pointing out the 2 EU distributor.

Since the case, heatsink, and Furutech are shipped from Conrad, Mark, and Davide, maybe it makes more sense to combine with just the Fuku's and PRPR's (which I am in possession for the moment) GB's.

For the 2K2013/2J313, a little more detail:
1. match means Vgs N-N and Vgs P-P.
2. Juma TF5 amp needs 2 x [2SK2013/2SJ313 triples] for 2 channels.
3. Juma TF5 line amp needs 2 x 2SK2013/2SJ313 for 2 channels.
4. BA-3 front end needs 2 x sSK2013/2SJ313 for 2 channels.
5. I will attempt to find close sets for the 2 channels but can't guaranteed.
6. I can't sell just N or just P, it has to be at least 1 pair of N-P.
7. Pricing is $4 per 2SK2013/2SJ313, e.g. One set of SK2013/2SJ313 triplets
is USD $12.
8. Anyone in this list can opt out any time, serving ordering is as following:

me, 2 sets of 2SK2013/2SJ313 triplets (Juma TF5),
4 sets of 2SK2013/2SJ313 (one stereo TF5 line amp,one stereo BA-3 FE). USD$40.
Smyslow, if possible, I will get you a set of 2SK2013/2SJ313 quad. If the yield
doesn't cooperate, then I may resort to 2 sets of doubles. USD$16
Heinz, 4 sets of 2SK2013/2SJ313 triplets. USD$48
Jameshill, 2 sets of 2SK2013/2SJ313 triplets. USD$24.

Thanks,
 
Status
Not open for further replies.
Top Bottom