1 channel of MiniA working - measurments
Finnaly I've manage to test atleast first half of my MiniA (L channel that is) ;)
So my findings:
- well, it works (at least in mono arrangement ;) )
- measured numbers are (the marks of elements are from A5 schematic in Service manual) :
o DC offset at output 28-30mV
o Voltage drop on R11 5,53V
o From Source of Q1 Q2 pair to gnd: 3,58V
o Q4 C-E: 4,94V
o Q5 C-E: 4,94V
o Z5: 9,2V (I guess my voltmeter ain't very precise)
o And on power source resistors in both cases 0,47V ΰ so I guess bias is currently set to around 1A or maybe even a bit less (since my resistor is made out of 2 parallel 1R ones). My goal is to get around 1,3A of bias ..
o D-S voltages on irfp240 are 15,1V
o BUT there is something I wonder I bought 9610 from a guy as allready matched pairs
Now in working circuitry I've measured VGS of them: Q1 = 3,56V, Q2 =
3,59V (hmm that's allready on the border of 30mV) and Q3 = 3,71V?!?!? (but this one draws a bit higher current and so VGS is higher, right?) Now I just wonder the guy who sold them to me matched them to 5,26V (in the case of these pair) so I guess he did it at much higher current but and now my Qusetion:
is this current/VGS ratio linear or not? Does this difference of Q1 and Q2 means they are not well (I've got another 8 »matched« pairs of them waiting for my projects) matched at all or just not matched for my demands?
I have wondered the same thing and can only assume that a matched pair is NOT a matched pair unless matched at the same current as the situation. I have had the same results as you, stabist. I have resorted to matching them 'in situ'. I don't solder them in real good until I swap in ones that match...
Well now after first excitement at 4am is gone ;) - I've readed articels abouut matching - and realized I'm right - if transistors are matched at such current that VGS is e.g. 5V - that doesn't mean they will be matched at the current with VGS 4V ...
Another way than yours mpmarino is simply add that additional resistor ...
BUT!! then again - how can I calculate enough precize value - because my DMV has got not enough digits to find out is it 25mV or 34mV difference ...
Btw - this resistor - it is put in series to source of the transistor with lower VGS - RIGHT??
I would consider power MOSFETs matched within 10mV Vgs a very good match (so 30mV is not so bad). Unlike FETs, bipolars are much easier to match so matching within 2 mV is usually not difficult to obtain.
Anyway all you need to measure Vgs (for matching purposes) is a DC source of about 9-12V, resistor and a voltmeter. To measure N-channel devices connect the source to 0V (GND), short the gate and drain and connect a resitor between the drain and + pole of the DC source (Vdd). At this point you measure MOSFET's Vgs that corresponds to the drain current (Vdd-Vgs)/R where R is the value of the used resistor (i.e. test Id is set by R). To test P-channel devices just swap the poles of the DC supply.
I am pretty sure Nelson advertised that matching was done 'in situ' on the later alephs. Personally it works for me, and I have found that it is more accurate and takes less time. When you get a pair that has low offset...your done! Easy as pie and no tedious measurements or wondering. My 2 pennies.....
I have not really learn how to match transistors but your method sound easier. Can you write out a few bullet points of what you meant by "in situ" and the procedure of how you go about achieving it?
I think you meant by poping in/out the transistors by trial and error and measuring it and see which one(s) will get the closest match, to within 10mv. Questions like .... , which pin terminals to measure, etc ...?
Yes, I did mean trial and error.. to a point (I guess you could go in circles for a while!). On the two alephs I have built (one a mini A, so the output fets being matches is irrelevant) I attempted using 'matched' 9610's only to find a little more DC offset at the output than I was happy with. I then simply focused on only 1 fet in the diff pair and kept swapping in a new fet until I was happy with the output DC offset. Maybe you would have to give up on that one and try swapping the other fet to make it balance. The fets I have are from the same lot and I started with quite a few, they were pretty close already. Seems to me you can have guaranteed good results if you have enough fets. Funny that the fets I ended up with in the amps were not matches from my 'matching session'. So much for my matching abilities! :xeye:
Stabist comment about measuring VGS throw me first. What does it mean by measuring VGS? Is that the voltage between pin G and S?
To make sure that I understand you. You are saying ... pop the 9610 in and measure ONLY the output DC offset? No need to measure anything else?
We want to have output DC offset to be less than 10mV? 5mV?
Thanks - alltough I have to admit I know this circutry - it's just the thing I'm a bit pissed because I've bought like exactly matched pairs from this guy - but now it seems that doesn't be so ...
And OK, I can live with that from the aspect of 9610 - BUT - I also bought 14 matched 240 ones for my A30 - and now I have doubts they are matched well ... And in the case of output transistors - I'm a bit more worried ... :((
Will have to test them myself I guess - so I just have to get some extra time ...
Yep VGS is G-S voltage ...
I think that this method mentioned by mpmarino is also a bit slow - I think you must wait for each measurment again like 10, 15min - so everything settles down and the amp worms up and all ...
Please understand me, I'm no expert and have very little experience with this circuit compared to some/most on this forum...just that I feel like I learn some by discussing things. Anything I say that's off I expect to get corrected, in fact I hope so!
Yes measuring vgs is simply turning the fet on by running current thru and measuring voltage gate to source. Nelson has a simple procedure somewhere on his passdiy site, and chip_mk described it a few posts ago.
I am saying that in my experience, vgs matching of the diff pair got things pretty darn close, but I still had maybe 20-30mv at the output (actually I guess that's OK but I wanted it better. I assume that unless every component(1/4 watt resistors around the 9610s..etc..etc) is perfectly matched you will have offset at the output. Finding the 'magic' pair worked for me. The way I understand it,the goal of matching the differential pair is to approach 0v offset. Please tell me if I'm an idiot:smash:
Stabist, your absolutely right. It is slow!
|All times are GMT. The time now is 12:49 PM.|
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2013 DragonByte Technologies Ltd.
Copyright ©1999-2013 diyAudio