Question on device matching

Status
Not open for further replies.
Retired diyAudio Moderator
Joined 2002
I know I have seen this before, but I searched for quite a while and could not find the post.

When I am matching my IRFP240 devices for the Aleph 2 that I am building, How long should I keep them with applied voltage? I was matching some today, and it seemed to keep on changing. What is the recommended time for stabilization? I know that I have seen either Nelson Pass, or Mark Finnis post about this before.

Also, what is the optimum current for matching the output components? 200mA? What have others used for matching input and output components (current wise).

One more thing, I have access to a new $30,000 Tektronics digital curve tracer here at Georgia Tech. Is there anything else that I can do with that that will allow for better matching of any of the stages, or should I just stick to the measuring the VGS method?

Thanks for all of your help.

--
Brian
gte619j@prism.gatech.edu
 
Brian,

I used a car battery charger at about 13V, and 16 ohms of resistance. THis gives you about 0.5A thru the IRPF240 which is what the idle current is. I didn't use a heatsink and the devices get quite warm but not particularly hot. I let them heat up for 5 minutes, at which time I took my Vgs reading. By 5 minutes they are good and stable.

On the curve tracer, someone else will have to help you out, I don't have the expertise there.
 
Brian,

I matched the IRFP240´s at 700mA (for an Aleph5). I put them on a small heatsink and waited till Vgs didn´t change for two minutes. Most devices were constant after 5 minutes, a few took 7-8 minutes.
I was able to get 9 sets of 3 out of 30 with a max diff in Vgs from 0,07V .
The min value was 4,25V, max. was 5,03.

william
 
I purchased 125 IRFP240s last week from digikey for an aleph 2 and they arrived today. I am here at Georgia Tech, and have access to a good deal of equipment at the ece department for matching.

My current plan is to use an hp voltage source, and run .5 ma through the irfp240, in the process described by Mr. Pass in the article on the website, amd measure it with a hp dmm. I figured that I would time them all for 5 minutes and record the voltage and move on to the next one.

I matched some IRFP044s this way at .25mA just to see how it would go, and it seemed to go alright. They did get quite hot, as I did not use a heatsink.

For matching the IRFP240s, do I need to put them on a heatsink? Will simply hooking them up for 5 minutes without a heatsink, running .5mA through them work properly?

--
Brian
gte619j@prism.gatech.edu
 
It's the Temperature

I suggest you use a heatsink. In fact, use the same heatsink you plan to mount them on in the final amplifier. More importantly, you'll want to make sure each device is tested the same way. The Vgs threshold voltage is rather temperature dependent.

I don't know how closely matched they need to be for this particular amplifier, but wait until the reading is stabilized (as the other posts say). Basically, what you are waiting for is the temperature to settle to a final value. The larger the heatsink, the longer you will wait. But remember, you want to match devices at or near how they will operate in the amplifier. And that means temperature, drain current and voltage (bias).

The expensive machine won't do any more for you than a cheap DVM.

jh
 
I purchased a smaller heatsink for this testing and plan on clamping the IRFP240 to the heatsink with a woodworking spring clamp that I purchased from Home Depot, and waiting 5 minutes, and recording the VGS value that it gives. I will also be sure to retest the first one, since it will be the time when the heatsinks first warms up. This should give an accurate measurement.

Assuming I have a powerful enough voltage source, do you think that there would be complications if I tested 2 at a time on the same heatsink? I have matched resistors, and I can place them similiarly on the heatsink. I am just looking to cut my matching time in half. I have 125 to test, and it would help. If I keep it constant and always test two at a time, and ignore the last one, does this seem alright?

--
Brian
gte619j@prism.gatech
 
Brian,

Watch out that the drains are connected to the metal backing of the transistors.... if that's OK, then go ahead, otherwise you'll have some smoke possibly.

On another note, you're biasing for 0.5A, not 0.5mA, right ?????

I think the heatsink is going to give you poor results. The reason is that as you're testing the transistors the sink is going to be getting hotter and hotter as time goes on until it reashes its equilibrium temp for the current you're testing to. I just tested the bare transistors and had no troubles - it's bad enough waiting for the transistors to heat up let alone waiting until a heatsink reaches its terminal temp...... Takes maybe 3 mins for bare transistors to heat up, then 2 more just to settle a bit.... so that 5 mins per device is plenty.
 
Wayne, you are correct, I meant to state 0.5A for the biasing. I ran the setup last night with the heatsink, and matching one device. I started it with leaving the first device on there for 30 minutes, then putting it at the end of the line and testing from the beginning for five minutes each. This seems to be a good method. They stabilized around 3 minutes, and I had them timed for 5 minutes each. I made sure to let the heatsink stabilize to room temperature before this. I think tonight that I will try it without the heatsink and see if I get similar results. I plan on grabbing 4 voltage source setups (there are two at each workstation in the senior lab), so I can match 4 at a time without heatsinks.

As for the device and the drains connecting the to heatsink, I took care of this by using an insulating wafer.

--
Brian
gte619j@prism.gatech.edu
 
results of initial matching

I ran my trial setup with some IRFP044's. I am going to run the matching for the 240's the same way.

15v, 0.5A for 5 minutes with no heatsink. The heatsink was causing problems. Here are my results from the 8 044's that I matched tonight:

# VGS
0 5.3285
1 5.3231
2 5.3345
3 5.3195
4 5.3095
5 5.3312
6 5.3042
7 5.3065

They were all from the same lot. I started off and put one of them on there for 30 minutes to get the resistors up to temp and get the test bench stabilized. I then put this one aside and tested my others. I left them all sit and equalize to room temp before all of this. I then started with 0 and went to 7, running each for exactly 5 minutes. After 5 minutes they got pretty hot and pretty much stabilized. These seem like good matching results, and I hope that the 240's come out this good. I will take a picture of the testing setup. I got some sockets for the output devices, so that I can easily swap them. The are resting on nothing, suspended in the air. The room had a constant temperature.

Does anyone need some matched 044's? I will sell them to anyone who needs some. I just got them to test the setup, since they were all I could find locally.

I am going to match 125 IRFP240's next week, if anyone needs sets of matched IRF240's, e-mail me. I should have plenty come next week.

--
Brian
gte619j@prism.gatech.edu


--
Brian
gte619j@prism.gatech.edu
 
Status
Not open for further replies.