[BoilermakerFan]

Everything arrived, but I did not get started over the weekend as planned. I had better offers for night time activities.

Starting the builds, a V1.0 and V1.2, both as BJT buffers. The roll call for testing DB pairs as it stands right now is:

2238/968
3422/1359
2344/1011
3421/1358
649A/669A
5171/1930
1226/690 (Obsolete - salvaged from Mid-Fi Garrard integrated receiver)
1913/1274
507/313

I've managed to misplace a few previously untested DB pairs, and a couple of well recognized boutique sets are also AWOL for now.

I'm using BC546C/556C small signal transistors, so I have 3 packs of the standard transistors matched by Tom up for sale if anyone needs them. I'll entertain equal value trades for caps, LEDs, etc.

 

[BoilermakerFan]

Quote:

Originally Posted by tomb /img/forum/go_quote.gif I went with 0.75A and it's worked fine for up to 150ma bias - would probably go higher, too. QB2L/R - ZVP3306A (P-channel MOSFET) QB3L/R - ZVN3306A (N-channel MOSFET) I bought ten each - remember that they must be matched according to current. EDIT: Also remember that they are reversed compared to the silkscreen on the PCB - the silkscreen still follows the BJT version for QB2 and QB3. Yes - it was your idea first to start looking at those. I hope those and the Zetex's (misspelled it earlier) work out.

OK, now I'm showing that I'm still a newbie, but can I use my el-Cheapo Harbor Freight meter to match the Zetex? How does the DGS pins of the MOSFET correspond to the EBC pins on a BJT transistor? I'm assuming the meter is low power so they wouldn't toast, but I'm also not willing to risk it.

Tom, I'll send you my N and P channel numbers for my 10 to compare to yours and see if we might be able to match up any extremes we have in out limited samples.

 

[luvdunhill]

Quote:

Originally Posted by BoilermakerFan /img/forum/go_quote.gif OK, now I'm showing that I'm still a newbie, but can I use my el-Cheapo Harbor Freight meter to match the Zetex? How does the DGS pins of the MOSFET correspond to the EBC pins on a BJT transistor? I'm assuming the meter is low power so they wouldn't toast, but I'm also not willing to risk it. Tom, I'll send you my N and P channel numbers for my 10 to compare to yours and see if we might be able to match up any extremes we have in out limited samples.

nope, you'll need a power supply and a resistor to measure the MOSFET Vgs. Check out AMB's site on how to match MOSFETs. I would argue the term match by current, but I understand what you're trying to get at.

The Zetex are good transistors, I have a bunch of them myself.

 

[BoilermakerFan]

Quote:

Originally Posted by luvdunhill /img/forum/go_quote.gif nope, you'll need a power supply and a resistor to measure the MOSFET Vgs. Check out AMB's site on how to match MOSFETs. I would argue the term match by current, but I understand what you're trying to get at. The Zetex are good transistors, I have a bunch of them myself.

Thanks. A similar circuit is also in Jones' book on building valve amps, but I thought it was for DMMs without HFE.

I may just keep the BC546C/BC556Cs as I have sets that are matched insanely close (every pair/quad are within 10. One full set is within 8 for the whole amp and most pairs are exact matches within the capabilities and error of my DMM) and I'm building the V1.2 prototype as a BJT, I'd save the Zetexes for the official MOSFET-MAX build.

 

[tomb]

Quote:

Originally Posted by BoilermakerFan /img/forum/go_quote.gif Thanks. A similar circuit is also in Jones' book on building valve amps, but I thought it was for DMMs without HFE. I may just keep the BC546C/BC556Cs as I have sets that are matched insanely close (every pair/quad are within 10. One full set is within 8 for the whole amp and most pairs are exact matches within the capabilities and error of my DMM) and I'm building the V1.2 prototype as a BJT, I'd save the Zetexes for the official MOSFET-MAX build.

Technically, it's matching for I[size=xx-small]DSS[/size]. This is done by measuring current in the circuit shown below, which puts the JFET into saturation at its I[size=xx-small]DSS[/size]:

An easy modification to the circuit shown above is to substitute a 100R resistor for the ammeter and then simply read voltage across the resistor with your DMM and convert to current. I have a little breadboard set up to do this. It's a bit more difficult than matching for HFE on BJT's, but it goes fairly quickly once you have everything set up.

 

[luvdunhill]

Quote:

Originally Posted by tomb /img/forum/go_quote.gif Technically, it's matching for I[size=xx-small]DSS[/size]

The ZVP3306A/ZVN3306A are MOSFETs, and as such you want to use a setup like this:

Ideally, the voltage and resistor will set the test jig as close to how the devices will be operating in the target circuit.

You'll want to match to around 100mV difference at a bare minimum, but 10mV is possible with a standard hand held DMM. If you're looking for any more resolution than that, you'll need a nice bench meter. The difficult lies in controlling the temperature, as this will be the largest source of error in your measurements. Either mount them to a large block of aluminum or use an enclosure of sorts if your target current is small enough to run the device with free air heatsinking. I use a thermocouple to ensure the measurements are taken plus or minus a few degrees Celsius.

In fact, there is a better way to match MOSFETs that involves using a current limiting bench supply and another adjustable supply between the gate and source, but it's a bit more complicated.

 

[Pars]

Quote:

Originally Posted by luvdunhill /img/forum/go_quote.gif <snip> In fact, there is a better way to match MOSFETs that involves using a current limiting bench supply and another adjustable supply between the gate and source, but it's a bit more complicated.

Have you got a pointer to this method? Thanks.

 

[luvdunhill]

Quote:

Originally Posted by Pars /img/forum/go_quote.gif Have you got a pointer to this method? Thanks.

Can't say there is a reference. At the time I was matching $10 MOSFETs and wanted to get them very close and discovered a way that is more accurate, but a bigger PITA. More accurate meaning that the matching process exactly duplicates the operating point in the circuit. This is the key really, and as I said the temperature of the device is the hardest element to control. Then again, my target current was around 2A. You really need a current limiting supply and a something to bias the device and adjust until you are drawing the desired current, then measure and record Vgs. You can use a 9v battery and a pot if you don't have another bench supply. Also you should model the target junction temperature. In my case this was 60C, and once this was met I took down the Vgs reading.

 

[BoilermakerFan]

Quote:

Originally Posted by luvdunhill /img/forum/go_quote.gif The ZVP3306A/ZVN3306A are MOSFETs, and as such you want to use a setup like this: Ideally, the voltage and resistor will set the test jig as close to how the devices will be operating in the target circuit. You'll want to match to around 100mV difference at a bare minimum, but 10mV is possible with a standard hand held DMM. If you're looking for any more resolution than that, you'll need a nice bench meter. The difficult lies in controlling the temperature, as this will be the largest source of error in your measurements. Either mount them to a large block of aluminum or use an enclosure of sorts if your target current is small enough to run the device with free air heatsinking. I use a thermocouple to ensure the measurements are taken plus or minus a few degrees Celsius. In fact, there is a better way to match MOSFETs that involves using a current limiting bench supply and another adjustable supply between the gate and source, but it's a bit more complicated.

Will the MOSFET test circuit also for JFETs or is Tom's JFET diagram required? I was going to protoboad the circuit with a potentiometer for the resistor to allow adjustment, use TO-220 and TO-92 transistor sockets in parallel, then and use a DPDT so I can flip a switch to do N-channel or P-channel with either TO-220 or TO-92 FETs. I could proto both on the same board with another DPDT in the center, but one is just easier than two...

My Sears DMM is acting up and the display only works when the case is fully opened which means it's warped and loses contact when it's tightened down on the sides. Picking up a Amprobe 37XR-A as a replacement later today...

 

[luvdunhill]

Quote:

Originally Posted by BoilermakerFan /img/forum/go_quote.gif Will the MOSFET test circuit also for JFETs or is Tom's JFET diagram required?

First, I don't really know if JFETs are used in the MOSFET-MAX to be honest, I was just answering your question about how to match MOSFETs. So, I'll leave that question to Tom

But the answer the above, I wouldn't recommend it. The resistor in the MOSFET matching jig needs to be able to dissipate a good amount of heat, depending on the matching current. Notice that AMB suggests a 5W resistor in the above diagram.

 

[tomb]

Quote:

Originally Posted by luvdunhill /img/forum/go_quote.gif First, I don't really know if JFETs are used in the MOSFET-MAX to be honest, I was just answering your question about how to match MOSFETs. So, I'll leave that question to Tom But the answer the above, I wouldn't recommend it. The resistor in the MOSFET matching jig needs to be able to dissipate a good amount of heat, depending on the matching current. Notice that AMB suggests a 5W resistor in the above diagram.

It's plenty for the currents used in the Xetex's and Toshiba SK's - and even PN4392's and 2N5486's. The Toshiba SK's average 7-10ma. That's 0.01W - plenty for a measly 'ole RN55. The PN4392's and 2N5486's are less than half as much.

No one said anything about matching the power MOSFETs. Even AMB says that's unnecessary and nowhere in the history of everything Millett MAX/MOSFET MAX/MiniMAX have we ever recommended matching anything of a power output nature. MOSFETs should be tested, which can be done holding the two probes of a DMM in your hand along with the MOSFET - nothing else.

P.S. They're all FETs, whether MOS or J, they're still being used for the same purpose (as a JFET) - the first diagrams I posted should work fine.

 

[BoilermakerFan]

Quote:

Originally Posted by tomb /img/forum/go_quote.gif It's plenty for the currents used in the Xetex's and Toshiba SK's - and even PN4392's and 2N5486's. The Toshiba SK's average 7-10ma. That's 0.01W - plenty for a measly 'ole RN55. The PN4392's and 2N5486's are less than half as much. No one said anything about matching the power MOSFETs. Even AMB says that's unnecessary and nowhere in the history of everything Millett MAX/MOSFET MAX/MiniMAX have we ever recommended matching anything of a power output nature. MOSFETs should be tested, which can be done holding the two probes of a DMM in your hand along with the MOSFET - nothing else. P.S. They're all FETs, whether MOS or J - the first diagrams I posted will work fine. In the signal input positions of a Diamond Buffer (QB2 and QB3), the I[size=xx-small]DSS[/size] is all you're interested in matching.

Thanks guys. I'll build one of them and test my FETs...

Picked up my new meter today too. Local, family-owned electronics shop even knocked $5 off the price!

Amprobe® Test Tools

 

[tomb]

Quote:

Originally Posted by BoilermakerFan /img/forum/go_quote.gif Thanks guys. I'll build one of them and test my FETs... Picked up my new meter today too. Local, family-owned electronics shop even knocked $5 off the price! Amprobe® Test Tools

The Xetex's test out at 65-70ma. So, you might want to use an RN60, instead. That's 1/2W, but even an RN55 would probably be OK - you don't leave the FETs stuck in there for very long.

In light of that, a 1000R resistor might be a better choice - these do seem to move more current than the others I mentioned.

 

[BoilermakerFan]

Quote:

Originally Posted by tomb /img/forum/go_quote.gif The Xetex's test out at 65-70ma. So, you might want to use an RN60, instead. That's 1/2W, but even an RN55 would probably be OK - you don't leave the FETs stuck in there for very long. In light of that, a 1000R resistor might be a better choice - these do seem to move more current than the others I mentioned.

RN55 are actually 1/2W, just derated to 1/4 for MIL-SPEC, but I have RN60s for other projects and a couple 2W Caddocks.

 

[tomb]

Quote:

Originally Posted by BoilermakerFan /img/forum/go_quote.gif RN55 are actually 1/2W, just derated to 1/4 for MIL-SPEC, but I have RN60s for other projects and a couple 2W Caddocks.

Actuall, it's 1/8 to 1/4 for RN55's and 1/4 to 1/2 for RN60's. Still - there's no reason you would want to leave these in the circuit that long anyway. It's quite similar to matching BJT's or even resistors - consistency in comparison is what's important, not absolute measurement. Pop the FET in the circuit, take a reading, then pull it out and pop the next one in, etc., etc.