Page 4 of 6

#46

Posted: Sat Oct 13, 2012 9:57 am
by simon
Nick wrote:That case and heatsinks are from

http://www.modushop.biz/ecommerce/cat143_l2.php?n=1
Out of interest Nick how do the chips attach to the heatsinks? Are they already drilled and tapped, or do they attach to the sides of the chassis and a good thermal bond from the chassis to the heatsinks is relied upon?

#47

Posted: Sat Oct 13, 2012 12:04 pm
by Nick
I drilled and tapped them, then used insulating pads under the devices. Check that the face is flat and not raised around the hole or it will cut through the insulator.

#48

Posted: Sat Oct 13, 2012 12:06 pm
by IslandPink
In the case of the Mosfets in the F4, you have to drill & tap M3 holes in the Heatsink, and use a fitting kit between the transistor and heatsink which uses either a mica washer ( shim ) plus some Zinc paste for heat conduction, or a mica washer coated in some heat-sensitive bonding layer that melts when the unit gets up to temperature and bonds the pieces together . I used the former, Nick used the latter .
It's one of the more fiddly aspects of the build, as you have to mark and drill the holes fairly accurately . I think I fitted the Mosfets to the heatsink first, with their little legs sticking up, then tweaked all the leg positions as I brought the PCB board up to them, pushed the board onto them, then soldered all the legs .
It's very very helpful to use at least the signal PCB boards that are available via DIY Audio posters. I used the Peter Daniel boards for the F4 , he's out in Canada .

#49

Posted: Sat Oct 13, 2012 12:57 pm
by simon
How easy is to drill and tap, it's not something I've ever done. Presumably you need a drill bit and a tap of some description, something like this?:

Drill the hole out first, then tap it by hand?

#50

Posted: Sat Oct 13, 2012 1:17 pm
by pre65
For M3 you would need to drill a 2.5mm hole, and then use the taper tap first, and the plug second.

Don't forget you will also need a tap wrench. :)

To be honest in "hard" aluminium you probably don't need any special lubricant, just feed the tap in a turn, then back off half a turn and keep going like that.

Goes without saying (but I will anyway) the holes need to be marked out accurately.

Tap wrench link.

http://www.ebay.co.uk/itm/TAP-WRENCH-M3 ... _598wt_699

#51

Posted: Sat Oct 13, 2012 2:20 pm
by colin.hepburn
Nick wrote:
colin.hepburn wrote:Is it absolutely necessary to use match transistor in the Nelson Pass F5
What are likely to be the consequences if unmatched transistors are used?
Most likely higher distortion. The setup equals the idle current through them so its mainly gain that needs to me matched.
AH well a bit of an overreaction there I think Steve with the word Bang :shock: :) as nick said maybe distortion could be the result

But then I have had an amp in the past a mission Cyrus one with unmatched Drive Transistors replaced and no bangs or notested distortion so maybe not so critical but mosfets could be a different story of course

Trouble with getting matched Transistors is most suppliers won't match then for you so you have to buy loads and match them yourself which I don't have the means to do that

Or you have to pay more to buy from someone who has already done the matching a bit of a pain and still no guarantee you have been sold matched pairs

#52

Posted: Sat Oct 13, 2012 2:35 pm
by pre65
No one said it was easy. :wink: :)

#53

Posted: Sat Oct 13, 2012 2:51 pm
by Cressy Snr
colin.hepburn wrote:
AH well a bit of an overreaction there I think Steve with the word Bang :shock: :)
Just kidding :lol:

#54

Posted: Sat Oct 13, 2012 3:19 pm
by Mike H
Re tapping a hole, beware snapping the tap in it if it's too tight. I.e. hole size has to be 'just right'.

Guess how I know :lol:

#55

Posted: Sat Oct 13, 2012 3:32 pm
by pre65
Mike H wrote:Re tapping a hole, beware snapping the tap in it if it's too tight. I.e. hole size has to be 'just right'.

Guess how I know :lol:
According to my Dormer Tap and die handbook the M3 minor thread diameter is 2.459mm so the recommended tapping drill of 2.5mm means it should not be "tight".

But as always, use your loaf when doing anything like this. :wink:

Also, if not drilling a through hole (ie blind hole) drill a tad deeper than you need (+25%) and clear out swarf after drilling, and after taper and plug taps. A pillar drill will most likely have an adjustable stop so you can drill consistent depth blind holes.

#56

Posted: Sat Oct 13, 2012 3:58 pm
by colin.hepburn
pre65 wrote:No one said it was easy. :wink: :)
Hi Philip
Yes this is the problems we all face isn't it I am afraid :(

but I am one of those who if it become a big problem to finding parts which are often over priced and or fake I just tend to give up on the project and move on to a another project just to start the same problems over a gain

There are F5 and JLH kits on eBay for example but don't think I would trust them to have good matched transistors :shock:
So short of contact Mr. Pass himself and buying from him I feel that's the only way to be sure you get the correct parts :)

#57

Posted: Sun Oct 14, 2012 5:12 pm
by Mike H
Interestingly I never bother too much about matching, I start from the basis that they're not, :D but if from the same batch, they might be 'close enough'.

It then makes me think about how best to offset any mismatch by applying the right sort of biasing.

If you think about it a lot of commercial circuits are like this, because it's the most practical for chucking circuits together on a production line.

#58

Posted: Sun Oct 14, 2012 6:05 pm
by colin.hepburn
What are we looking for to match transistors is it to do with the hfe gain and how would we go about the matching process i.e. what is need to do this
I have see a transistor tester on the DMM meter but it couldn't be that simple could it :o the other problem I would have is cost with haven't to buy a hole box of transistors just to build one amp :shock:

#59

Posted: Mon Oct 15, 2012 1:14 pm
by Mike H
This is the problem! :D

I think nowadays accuracy is more predictable, like if it says on the data sheet Hfe between 150 & 200, for example, it better be that else "not as described / fit for purpose"

There are 2 ways of measuring it, the simplest 'static' bias method needs 2 multimeters, one for base current, uA (or whatever), second for collector current, mA (or whatever).

Then some sort of power supply WITH current limiting preferably! A PP3 battery with a 1k resistor in series can work quite well.

Then you need a largish value potentiometer to apply the base current, and measured with meter #1, then read the collector current on meter #2.

The gain is Ic / Ib.

The AC measuring method is much more complex, I remember a project for this in Practical Electronics years ago.

Even more fun is measuring the reverse cb leakage current of germanium transistors; with bias pot at zero, flows backwards through meter #1. :lol:

#60

Posted: Mon Oct 15, 2012 2:23 pm
by Nick
Remember they are mosfets not bipolar, Its the input pair that is most important, the output pair has bias setting pots. Simple way to match them is to treat them as current sources, and measure the current.

www.passdiy.com/pdf/matching.pdf

The F5's I have built have used Peters PCB's and semiconductor sets from someone in Germany via DiyAudio, I never mentioned eBay.

If you want to match your own, thats fine, but it will work out cheaper to by them matched if you only plan on making one amp. Remember that if you buy 20 pairs of devices, you can't assume you are going to get 20 matched pairs.

If you want I will find you the chap I bought the devices from.