F5m kit

Jump to Latest
#2,442
For non-monoblock, do the 140/9140's need to match only their compliments on 1 amp board, or do they also need to match the second amp board for the other channel?

I originally bought 2 F5m kits. I built one, but early on in testing my build with only 1 amp board connected to power, I thought I saw smoke and smelled burning and quickly turned the amp off. ( I had done and goofed with 2 140's instead of 140/9140. ) To be sure I was working with undamaged parts, I removed the amp board ( only had 1 powered for initial test ) and soldered up a board and components from my second F5m kit and used that one for my final build. It seem ok but not sure if I should check to make sure MOSFETS for both channels are 'matched'. I also think I may have used JFETS/MOSFETs mixed from my second kit. So in short, I may have mixed up JFETS/MOSFETS from 2 different kits.

Doing quick forensics, I think one of the 22ohm 2 watt resistors has a black hole in it and may have smoked.

To complicate things, I re-ordered new 140/9140's and other parts from mouser but not sure if I should check for matched pairs before building my second amp. Also, I could not get TOSH 2SJ74 and TOSH 2SK170 and am worried those may have been damaged.

I want to make sure I get this right before building my second F5m or modifying them to monoblocks.
 
#2,443
In Nelson's projects, the matching requirement for the input JFET P & N pairs are Idss within +/- 1mA. Matched JFETs purchased from the diyAudio Store will typically be much tighter than this, like within .2 mA. The output power MOSFETs need to match if same sex devices are being used in parallel, such as amplifiers like the F4 (3 in parallel), or the Aleph J (2 in parallel), and so on... These need to be matched for VGS so that the current (Iq) is shared equally, without one device "current hogging", in other words we don't want one device doing all the work. The "N's" do not need to match the "P's" for the outputs.
 
#2,445
Hello birdbox,

I have another opinion. Every amp benefits from matched FETs. In this circuit from Nelson Pass it works, that the FETs
don't have to be matched from the N-channel to the P-channel -FETs (adjustable over trimpots R8 / R9). Or let's say - not so closely matched.
I am pretty sure, that the MosFETs in the kits are matched into groups for Vgs. But I can't know for sure. How closely matched?
Cheers
Dirk
p.s.: a relaxed Christmas...
and always good sound 'under the tree' 😉
 
  • Like
Reactions: barossi
#2,447
You can make a simple jig to measure the mosfets and the JFETs. You can even use the 270ohm resistors to measure the mosfets at about a few watts of power if you like using the power supply on your amp.

The JFETs IDSS can be measured with a 12v supply and a single 100 ohm resistor. Well, since you would only be measuring to confirm matching, you don't even need exactly a 12v supply or exactly a 100 ohm resistor.

How many mosfets did you get?
 
#2,448
Hello birdbox,

this depends on a few factors. For example gain variations between N-channel and P-channel devices (if we are looking at a complementary amp stage) will result in different amounts of distortion (2nd / 3rd order harmonic distortion). But there is much more...
You can adjust or 'equalize' some of this behaviour with the choice of the bias-resistors.
Most important is matching for paralleled MosFETs in outputstages or differential inputstages (more often JFets / BJTs).
Some good articles:
https://www.firstwatt.com/wp-content/uploads/2023/12/art_mos_test.pdf
https://www.firstwatt.com/wp-content/uploads/2023/12/art_matching.pdf
Cheers
Dirk
 
  • Thank You
Reactions: birdbox
#2,449
@Mikerodrig27 I only got a couple of each. I guess my next step is to test the JFETs/MOSFETs i have not used yet to see if they are matched. ( if so, than I can assume the ones I used in my build are also matched )

@cubicincher thanks for those links. I'll see if I can build a test jig from those docs so I can ensure everything is matched.

If anyone posts a schematic for an all-purpose test jig, I can see about spinning up a PCB. ( could use the 'all test' in the art_mos_test.pdf? ) Don't want to re-invent the wheel but would be good to have a jig with proper test connection headers for the jfets/mosfets. I have one but it's for smaller footprint transistors and needs an oScope to view matches ( for synths )
 
Last edited:
#2,450
Hello roboDNA,

ther is a schematic on page 6 of this article from Nelson Pass (see above):
https://www.firstwatt.com/wp-content/uploads/2023/12/art_mos_test.pdf

I use a PEAK ATLAS DCA75 PRO for matching and curve tracing transistors. But its maximum voltage over the DUT (device under test) is limited to 12 V max. But I also got very good results with MosFETs running at higher voltages.
https://www.peakelec.co.uk/acatalog/dca75-dca-pro-semiconductor-analyser.html#SID=24
and if you want to test smd-JFets/BJTs (SOT-23) - this adapter is worth a lot:
https://www.peakelec.co.uk/acatalog/dca75-dca-pro-semiconductor-analyser.html#SID=24
But if your problem is a fried set of output-MosFETs (from the F5M - kit), then I would get in contact with the diyAudio-shop...?
Cheers
Dirk 😉
 
  • Like
Reactions: roboDNA
#2,451
@cubicincher good point, I can just order more matched sets from diyAudioshop instead of matching them myself. It will be quicker/cheaper for me just to order 4 new sets and replace the ones in my first build. ( considering the few times I'll need matched parts, it's not worth making the jig )

I can also take into account that I'll need matched sets for both channels of 'monoblock parallel' config but still need to learn about this. My amp is sounding great and the bass coming out of the super lintons is excellent. Still need to increase the gain as I have it at .700V now and heatsink temp is around 50c.
 
#2,452
If you truly just need more voltage gain, that can be provided by something like the "Iron Pre" preamplifier project that's well regarded around here. Increasing the Iq of your F5m output stage won't increase the gain. As Iq is increased, THD improves but you quickly reach a point of diminishing returns and thermal limitations. It sounds like you have yours set about where it needs to be. I like monoblocks for a few reasons... certainly the main one being the ability to transport one heavy / chunky (but still somewhat compact) Class A channel up and down a flight of stairs compared to two sharing a single large chassis. I build and modify these projects in a spare bedroom on the 2nd floor of my home.
 
  • Thank You
Reactions: roboDNA
#2,453
roboDNA said:
@cubicincher good point, I can just order more matched sets from diyAudioshop instead of matching them myself. It will be quicker/cheaper for me just to order 4 new sets and replace the ones in my first build. ( considering the few times I'll need matched parts, it's not worth making the jig )

I can also take into account that I'll need matched sets for both channels of 'monoblock parallel' config but still need to learn about this. My amp is sounding great and the bass coming out of the super lintons is excellent. Still need to increase the gain as I have it at .700V now and heatsink temp is around 50c.
Click to expand...
No more sets and mosfets avaible!
 
#2,455
Just completed a F5M with 2 x 200W trannies dual mono configuration. The chassis has 400x150x40mm heat sinks and I bias the board at 1.8A (after stabilised for 2 hours.) The heat sink temp at the FETs is 47C and 42~43C at other portions. The amp plays beautifully and is the most resolving one among all amps I owned, both solid state and tube. Since no one mentions using this bias current for F5 series of amps, I wonder if there is long term issue. If there is no long term issue, should I bias it to say 2A to have a even larger Class A envelope?
 
  • Like
Reactions: duder1982
You must log in or register to reply here.
Top Bottom