L'Amp: A simple SIT Amp

@protos

here you can see the safe area for the Sony Mike calculated, it is the zone below the bend doted red line.....
s2k82b.jpg



http://www.diyaudio.com/forums/diyaudio-com-articles/200460-l-amp-simple-sit-amp-part-1-a.html

is the complete article...

sorry the picture got a bit big.....and I do not know why.... first time such a monster picture....
Ok it does not look linear at high voltages.
However I tested one kd-33 with my lab ps as ccs and with source 1ohm 2amp -2v bias I was only getting like 6v Vds-only half what the graph suggests.
 
Very bad news.
The 8 kd-33 I bought from Acronman-circuitdiy have huge variations.
I tested with 3ohm source resistor and lab ps as ccs set at 1.2A for 30 secs (clamp to heatsink not ideal). 47K gate to ground.
The variation in Vds was peppered from 3.75V to 11.5V.
That is more than 300% variation!!!
There are a couple that are close but I am now not sure if they are all rejects and should not be used at all.
Should I ask for refund and try and buy from somewhere else or is this normal?
It is a lot of money for rejects!
 
I did not get that far yet. I was curious why my vfets were not operating according to Michaels graph. So I did a basic "matching" circuit test on the fly to see how close they are.
If the variation is so great it means the gain between two very different umatcheable transistors is unequal or am I analyzing incorrectly?
Or do I just try and match operating points by varying Vgs?
 
Last edited:
Pass DIY Apprentice
Joined 2001
Paid Member
A particular Vgs isn't especially important unless you plan to build the self-biased version. In that case, your selection is more narrow. Since we have no definitive information on the various grades we have to improvise and adapt.

I suggest trying the version with the negative bias supply posted near the begining of the thread. Shoot for the same operating point (Vds and Id) and don't worry so much about the actual Vgs. Then look at the gain and distortion (or listen) as NP suggests.
 
Very bad news.
The 8 kd-33 I bought from Acronman-circuitdiy have huge variations.
I tested with 3ohm source resistor and lab ps as ccs set at 1.2A for 30 secs (clamp to heatsink not ideal). 47K gate to ground.
The variation in Vds was peppered from 3.75V to 11.5V.
That is more than 300% variation!!!
There are a couple that are close but I am now not sure if they are all rejects and should not be used at all.
Should I ask for refund and try and buy from somewhere else or is this normal?
It is a lot of money for rejects!


you must hve something wrong, i have put same batch of acronman KD33 and have between 4V and 4,2 v with 2,5 ohm source . look at your source , drain resistor resistor and power supply it must be there ( i use 25K gate to ground) ..

good luck:)
 
I checked each transistor twice under the same conditions. The results were the same. There were a couple at around 3.75-3.85 vgs , a couple around 4.20-4.40 then it was 6 , 8, 9.5, 11.5.
I checked also with a higher bias setting and it seems the first four have the same gain i.e. the increase in Vds between one bias setting and the other higher bias were similar between the four. That is good. The other four seem to be of a higher gain but I need to test them more because they were overheating because of the high dissipation(high VdsXhigh bias) at the higher bias setting and so I could not get more stable measurements at the higher setting.
I must say it is a bit of a hassle trying to find ways to temporarily test TO-3 units on a heatsink compared to the to-247 type I am used to.
I guess the gain characteristics are more important than small differences in Vgs which means the curves are slightly shifted in parallel in the graph but still remain the same relative to one another however when there are huge differences in Vgs of more than double between fets then even the gain might suffer. In any case one will need a bias supply for sure unless you want a different Power supply for each channel.
Still having played with tubes which have obvious variations but which are still far far closer than what I am seeing with these fets.
I am surprised nobody had such variations in this forum since we mostly bought from the same source.
 
Last edited:
Protos, I am using the F3 (or more correctly, ZEN V9) without feedback, with "ordinary" CCS-no Aleph modulation), driving a pair of Visaton B200 above 150 Hz on Open baffle. If you can't do L'amp, play with the F3, its a heck of a good amp :)

I am thinking of either the F3 or the L'amp since I have both transistors. I am still trying to get some feedback from anyone who has heard both. I can do L'amp of course- it is such a simple build after all. In the end I will probably do both. Just that I would like to find some F3 boards to make life easier.
 
Member
Joined 2009
Paid Member
I am surprised nobody had such variations in this forum since we mostly bought from the same source.

maybe nobody checked so many SITs like you....:D

most of us bought only one or two pairs and speaking only for me, I did not check before building like you the values separate ....

I only remarked that I had an abbreviation of about 0,5V of Vgs compared with Mikes values to get the same Vds and current for the SIT....
and because the Vgs difference was not so big I did not care much for the difference..

For F3 pcb look here...

http://www.diyaudio.com/forums/group-buys/172257-reissue-f3-f5-pcbs-group-buy-8.html

and contact suds... and be patient.....:D
 
Last edited:
maybe nobody checked so many SITs like you....:D

:) at first i have ordered only 2 2SK82 , in fact ( thanks to michael and god N.P) it"s first a no-brainer amplifier. sounds good at 1.6amp 19V with inductor or at 14V 2.4amp with res load
in my system i could hear more difference changing source resistor type and power supply Cap or source bypass cap than changing bias... perhaps i'm still an audiofool:spin:.... sure i'am
 
Member
Joined 2009
Paid Member
Hi SIT-Men,

I am onboard again, my experimental assembly is working.....

so my normal amp can work parallel for comparison (BJ2)....
and no more destroying the case by the 120th change...

look my newest modification of "nearly" CCS (tamed Mufollower) and my last circuit change.

I simply put together two pictures from Mike, the upper part is his suggestion for a CCS with SJEP120R100,
the lower part is Mike´s normal suggestion for the SIT.....

the changes for the Mu are the drain resistor of 0R3 for the SIT and the 500R pot going to the 1000uF going to the drain of the Semisouth.

With the pot I could make at a certain position the k2 dominant for for lower and higher volume levels, followed by the k3 in a constant distance.
Without the pot the k3 got very dominant putting the volume up....

So it has the Zen factor!

Next experiment will be the inductor solution.....
 

Attachments

  • SIT MU2.jpg
    SIT MU2.jpg
    54.7 KB · Views: 617
  • 106_0108.JPG
    106_0108.JPG
    332.7 KB · Views: 616