Planning to change the MOSFETS on my faulty Aleph J. Looking into the option suggested by Extreme_Boky to buy a tube (25 MOSFETS) from the same batch from Mouser. And even if Extreme_Boky said they would be quite well matched, I would still like to make sure and measure them myself to check and choose.
Could anybody refer me to some good thread of the procedure to match MOSFETS, or perhaps exlain it right here what jig should I make, what power supply voltage I need and what to measure exactly? Vgs? I've never done that before, but I would like to try and learn from that.
Could anybody refer me to some good thread of the procedure to match MOSFETS, or perhaps exlain it right here what jig should I make, what power supply voltage I need and what to measure exactly? Vgs? I've never done that before, but I would like to try and learn from that.
I agree you should measure and match your parts, even if they come from the same
full tube.
You can start here:
https://firstwatt.com/pdf/art_matching.pdf
This is a nice simple procedure and gets you much of the way.
full tube.
You can start here:
https://firstwatt.com/pdf/art_matching.pdf
This is a nice simple procedure and gets you much of the way.
Yes, check them - definitely, as I did for my Aleph J. The fact was - they were all matched really well.... I cherry-picked 4 pairs for my build. The probability you'll get at least 4 tightly matched pares, is quite high if you get a tube from the same batch
.
Last edited:
we made consensus eons ago, lower third line is best
now I'm too old to dispute that
Well, to somehow explain this for others.... in a real heat-exchanger, flows go in opposite direction to maximise heat-exchange... (having continiously the biggest temperature difference gives the biggest flow, is like potential and electric current) interesting here is you also want your simiconductors to be as cool as possible besides the lowest average temperature of your sink and that is not at the top of the heatsink. So there is an optimumat a lower position, question is how quickly heat conducts through aluminium vs. how quickly it is carried away from the sink in the rising air... Not willing to restart that discussion, the best spot will be indeed 'somewhere' in the lower half of the sink
and yes it does make a difference of several degrees for your precious parts
damn greenhorns!
unbolt those heatsinks, and rotate them for 180deg
you can do that even without undoing wires, carefully
then you'll have your Fugly!, if you upload some pics of higher resolution ...... for us Geezers
edit: or just exchange places of top and bottom lid ..........or mount set of feet on top lid, so you can rotate it by wish
that's my newest invention - how to correct amp's phase and also using it as Peltier element
Hmm.. feet on the top.. Now you may be onto something! I'll look into swapping them around 180 deg today. The signal wire, and LS out wires should survive a surgical procedure like this, but my power wires are "cut to fit" and I don't think they will make the run to PSU if I swap them around.. Plenty of wire on hand (Thanks Jim for making me order a lot of wire!) so I can just remove power wires and and re do them. I'll take some temp measurements at repeatable points on the heat sinks before and after for fun!
Dazzled and confused..
Very nice boards and layout, even with those heatsink rotations
Thanks for the kind words.. pesky rotations under consideration!
Thanks Dennis and Boky,
I realize I've read that Papa's article some time ago. Still some questions:
At what current are you testing IRFP240?
Nelson refered to 170mA for output devices, but that's probably because that particular amp was using 6 (or 12) output MOSFETS in paraller, if I'm not mistaking. Testing close to the real 400mV/0.47ohm=851mA for Aleph J would be too much, I think, and would require DUT be mounted on the heatsink. So, what's good enough current for matching?
Another thing. If I do not have a proper regulated power supply, would it be good to use just some unregulated 15VDC wallwart laying arround? Or using any other voltage with adequatelly selected resistor able to disipate required power?
Sorry for asking simple questions, but this is the first time I think about the matching.
I realize I've read that Papa's article some time ago. Still some questions:
At what current are you testing IRFP240?
Nelson refered to 170mA for output devices, but that's probably because that particular amp was using 6 (or 12) output MOSFETS in paraller, if I'm not mistaking. Testing close to the real 400mV/0.47ohm=851mA for Aleph J would be too much, I think, and would require DUT be mounted on the heatsink. So, what's good enough current for matching?
Another thing. If I do not have a proper regulated power supply, would it be good to use just some unregulated 15VDC wallwart laying arround? Or using any other voltage with adequatelly selected resistor able to disipate required power?
Sorry for asking simple questions, but this is the first time I think about the matching.
Last edited:
damn greenhorns!
unbolt those heatsinks, and rotate them for 180deg
you can do that even without undoing wires, carefully
then you'll have your Fugly!, if you upload some pics of higher resolution ...... for us Geezers
edit: or just exchange places of top and bottom lid ..........or mount set of feet on top lid, so you can rotate it by wish
that's my newest invention - how to correct amp's phase and also using it as Peltier element
AJ boards are now flipped! Not to difficult to do, thankfully I had enough slack in the power wires that I didn't need to disconnect them. I did desolder the signal wire, and LS Out/Ground to make things easier.
Also moved thermistors up and out of the way of the transformer primary wires, didn't realize they got so hot! Shortened up a few other wires mounted LED's into front face, then after getting everything buttoned up, I plugged it into a circuit with GFI and blew my 2A fuse.. boo.. My amp will be quiet for a few days while I order a few spares!
Better pics as requested, thanks for the suggestions all!
Attachments
the heat dissipation at higher currents is a problem.... you need a temp. controlled environment, large heatsink.... the BEST is to try and do it... you'll understand
Last edited:
I agree, Extreme_Boky - learning from your own mistakes is the most valuable (and I will go and try!), but learning from others is usually cheaper. That's why I ask at what current it makes sense to do the matching at. Must one go to all the troubles of the temp controlled environment with the current close to the real working conditions, or matching the MOSFETS at some lower current that does not require heatsinks still makes sense because their Vgs would still be very much comparable to the real conditions?
You can not make mistake, especially if you get a tube of MOSFETS from Mouser to start with
The matching is really more like checking of Vgs at certain currents and tempretures. MOSFETs are positive temperature coefficient devices, so you'll have to take that into consideration. Your measurements will be a function of the room temperature as well.
You could decide to check the transconductance at higher currents... but in the case of Aleph J (2 paralleled devices only), I do not think it is necessary. Other's may disagree... I didn't do it -> and I get only a few millivolts' of deviation in measurements, across all 4 resistors. However, I did match the 0.47 resistors, out of a bag of 50.
So, the best is to try... start with what Nelson suggested (170 -200mA), and then play a bit because you'll be learning... rather than making mistakes.
Good luck
The matching is really more like checking of Vgs at certain currents and tempretures. MOSFETs are positive temperature coefficient devices, so you'll have to take that into consideration. Your measurements will be a function of the room temperature
as well.You could decide to check the transconductance at higher currents... but in the case of Aleph J (2 paralleled devices only), I do not think it is necessary. Other's may disagree... I didn't do it -> and I get only a few millivolts' of deviation in measurements, across all 4 resistors. However, I did match the 0.47 resistors, out of a bag of 50.
So, the best is to try... start with what Nelson suggested (170 -200mA), and then play a bit because you'll be learning... rather than making mistakes.
Good luck
Last edited:
peppennino,
To be honest, Zen, 6L6, or pFarrell will be best to answer that question. From what I gather it's a simpler board with less options and confusion (Jumpers and choices from what I recall 6L6 telling me). Obviously, I haven't built the current store version so my ability to compare is pretty limited.
To be honest, Zen, 6L6, or pFarrell will be best to answer that question. From what I gather it's a simpler board with less options and confusion (Jumpers and choices from what I recall 6L6 telling me). Obviously, I haven't built the current store version so my ability to compare is pretty limited.
Last edited: