Jump to content

The ultimate DIY? A Stax SRM-T2!


Recommended Posts

The board I will be building has 2SC3381. I don't have the board in hand yet.

Looking at the 2SC3381 and MPSA06 pinout, one should be able to mount the 2 MPSA06 into the 2SC3381 position without twisting the pins by positioning them in different orientation therefore not possible to bond the two MPSA06 together. 

Edited by mwl168
Link to comment
Share on other sites

Nice balance servo. 

Here is an old T2 board with small boards with HN4C51J replacing 2SC3381.


By the way Kerry, what kind of pins do you use with your small daughter boards?

Edited by JoaMat
  • Like 1
Link to comment
Share on other sites

The pins are from Mill-Max:

0995-0-00-15-00-00-03-0 and 6857-0-15-15-47-27-10-0

The Mill-Max site has a decent configurator for pins and quite a lot of options.

You might also consider these:  3114-2-00-15-00-00-08-0

I use them with standard dip sockets.

 This is my delay module



Edited by Kerry
  • Like 1
Link to comment
Share on other sites

Thank you. Them pins looks very interesting indeed.

Regarding my current project. Goal with it was to build a T2 on an original board with all modern components. After a visit in the ditch I returned to the narrow road. Rebuilt the batteries to original LED version on the board. Had to change some resistors as I reduced voltages - battery is 640V instead of the original 740V. Furthermore the change to fjpf2145 instead of 2sc3675 called for a Darlington par out of pair of fjpf2145 in batteries.

 Very easy to make a Darington pair on the original board. Yellow Teflon wire from base to a via on the board – perfect and the via has a perfect diameter.

 Some components (ksa1220, ksc2690,ksc1008 and j112) have different pin out as the original ones so I twisted their legs to fit the board. Thanks Michael (mwl168) for how to bend legs on the ksc1008.

 Bottom line: I believe I’ve built a decent version of DIY T2 with all modern components. I’m satisfied and am about to make myself not so sober (it’s Friday evening).

 P.S. some problem with one channel, unwanted noise and hum. Happens now and then if you have  T2s… 

Edited by JoaMat
  • Like 6
Link to comment
Share on other sites

  • 3 months later...

Hi All!!

Short story:

Nice to see the T2-project is still very much alive! I have a LOT of reading to do .....
Is there any major upgrade since 2015 for T2 (for original board run v1.0)?
- Why the change for R24+31??  

Long story:
I have been away from my T2, for some years - way too long ...
(due to an insane waterleak in the house 64m3 ... following extreme problems with insurance, 4 re-housings(!), rebuilding everything. But things are good now, and the house is very nicely rebuild.)


Meanwhile one of the trafos in T2 failed, - and is now replaced. 
I build new transistor- and tube-testers, and everything tested fine, so i dont know the cause :( Could be heat though ...
Anyways - i assembled and tested everything, and T2 is now online again, and playing like a dream. (No noise at all).

I really forgot, just how insanely nice this amp sounds! :dance:


/Mr.Raven - aka Pedefede
(had to shift username, because i forgot pwd, and my former email-domain is ancient, and no longer maintained.)

Link to comment
Share on other sites

  • 3 weeks later...

Hi all. Maybe a stupid question, sorry, but what about using Maida 21ct voltage regulators for the power supply ? 

It look like it have a nice PSRR, noise (20uV at 400v) and impedance. It could be confugured for voltages up to 525v and it could be used for negative voltages too (with a separate wind). What are the disadvantages of this regulator in comparison with original from Kevin ?

I am asking, because I already have 6 boards 🙂  

Thanks in advance !


  • Confused 1
Link to comment
Share on other sites

It might be a little premature but I think I now have a working T2 using a modified version of joamat's staxt2nc3fdh7 amp files and kgsshv dc supplies....

Initial startup looks good. Channels started up seperately. Heater transformer running at full voltage. High voltage transformer on a variac. virtual batteries within 100mV of 740V before any adjustment. DC balance between + and - sides of a channel <1V without any adjustment, all 3 leds lit on both channels.

once my hands stop shaking I will take a few photos... I have been dreading switch on...  in the meantime here's a few build photos of a finished amp channel top and bottom, the amp and psu

I modified the delay circuit on one of the psu boards to use a relay instead of the now unavailable solid state relay.

I modified the amp boards so that

1. they fit into a 400mm deep chassis, hifi 2000 dissipante chassis.

2. they have screw terminals for all connectors on the top.. (less neat but easier for testing and rebuilding)

3. all resistors of 1W or more go on the top of the pcb with a drill hole under for increased airflow

4. groundpane is between 1.1mm and 1.3mm from any solder point

5. pcb is cut in half so each channel can be built, tested and mounted independently,

6. solder points for pot and input jacks removed and replaced with screw terminals.

7. silkscreen has instructions for adjustment of pots

8. mpsw06 transistors replaced with mpsa06

9. small standalone heatsink for Q34 FJPF2145 in the 20mA current source darlington pair has drill holes in the pcb by the fins for enhanced airflow.

10 all components have schematic numbers included in the silkscreen

11 leds and adjustment pots mounted on top side of pcb like the original T2

12 removed the bias input and 5.1M resistor (since the kgsshv psu already has the resistor in the bias section)

I will be happy to release the gerber files of my modification once I am certain the amp is reliable

For the build:

wima film capacitors, mixture of koa cm1/2 and xicon 273 series 1/2W resistors, for higher wattage vishay pr02 and pr03

All resistors, diodes, zeners raised from the pcb.

pcb 2mm with 2oz traces.

All 1/2W resistors matched to 0.1% or better at 1khz on a good LCR meter between + and - sections of an amp board AND between left and right channels.

all non psu and psu decoupling caps matched to better than 1% at 1KHz on a LCR meter between + and - sections of an amp channel AND between left and right channels. (except for the pF caps which are just too small to measure accurately).

all leds and mpsa06 transistors matched on a dca75 curve tracer between + and - sections of an amp board AND between left and right channels.

all zener diodes matched to within 1% or better between + and - sections of an amp board AND between left and right channels.

separate umbilical cords for AC heater power and DC voltages

all internal wiring 1KV silicon all signal wires will be cardas chassis wire

volume pot tkd 2500 series

left to do

install rca signal wiring, volume pot and stax output socket

test with signal generator and measure distortion etc

collapse from T2 build anxiety

drill top of case for valve sockets

clean up wiring

pray to the god(s) of electronics

sort out grounds on psu

post moar pictures











Here is the schematic for the staxt2nc3fdh7 NOTE this is my reverse engineer based on the pcb and the original T2 schematic and has not been checked by anyone.  ( as far as possible component labels reflect the original T2 component labels. In the case of 2 components replacing a single component in the original T2 the components are now labeled Xpt1 and Xp2. In the case of additional components no present in the original T2 they start with the number 9XX. In addition both batteries are shown and the the components labeled with a Xp for the O+ side battery and Xn for the O- side battery.

Note the 140K resistors in the virtual batteries dissipate 0.423W each and I found xicon 1/2W 273 series discoloured and drifted by about 1% after about 6 months heavy use when placed on the underside of the pcb. So you may want to think about 1W resistors or place the resistor on the top side of the pcb.



psu low voltage, bias, +500V -500V and hv delay  reverse engineered from pcb and original schematic. not checked by anyone.



psu +250V, -260V and -560V reverse engineered from original schematic and pcb not checked


Edited by jamesmking
  • Like 12
Link to comment
Share on other sites

On 9/9/2020 at 6:30 AM, G600 said:


Listen to it, and post some more pics

listen to it? if you check my todo list nowhere does it say listen to it... measure it yes... listen to it? why would I do that? 🙃

latest update, internal signal wiring done (silicone not cardas - I need to buy more cardas chassis wire) and front panel completed. Both channels powered up at the same time for the first time. No drama so far. The heatsinks around the el34s runs a bit hotter than joamats mini T2.  I not done any temperature checks look good. All transistors seem ok and so do the high power resistors. .6922 heater voltage is a little high at 6.35V so I will need to add some dropper resistors. All other voltages are spot on. El34 heaters are at 6.1V which is about the same level as my mini T2. I found my first issue, the DC balance between  + and - halves of a channel is fine <50mv BUT the DC offset to ground started at -85V and even adjusting the offset pot all the way I only got it down to -33V. Exactly the same on the other channel due to close matching of components and matched valves. So I will need to replace the dc offset adjustment trimmers with something 2-3x the range. Anyway I know you have been waiting for the money shot: glowing leds and heaters so enjoy:


Im currently going for an extended power on idle test and will be periodically checking voltages and temperatures on both the amp board and psu. If that passes it will be time to feed it some test signals... I hoping for more detail in the square waves, more dynamics in the triangle waves and smooth fatigue free sine waves.

Here are the first measurements. Distortion 1Khz input Total Harmonic Distortion to 49Khz is around 0.0063% unweighted (average of 20 measurements) 2nd harmonic is around -90db and third slightly higher at -87db. The other harmonics are very low all less than -105db. total harmonic distortion plus noise is better than 0.018% (im not using shielded cables for this test AND the amp and psu do not have top lids and the cases are not grounded so the amplifier is almost certainly has better thd+n than I am measuring in this setup. At around 575V rms 5th harmonic starts to rise and at 600Vrms output all harmonis increase and odd order are higher than even order harmonics. NOTE this measurements where taken with around 35V DC offset so it is possible clipping may occur a little later with less offset...

Noise, THD, THD+N and level of distortion harmonics at 575V RMS, 1Khz output:575.thumb.jpg.0cb9a8a8dbba5bbe3dd40b2f2b447ecc.jpg

Noise, THD, THD+N and level of distortion harmonics at 600V RMS output 1Khz: The THD up to 49Khz is still less than 0.05%600.thumb.jpg.5e752f6ef473b7e768ac6054e14fbd3b.jpg

The amp runs quite a bit hotter than the mini T2, heatsinks near the EL34s run at 40C and power consumption is close to 205W. The mini T2 is around 153W.  The power supply is also is less efficient than the golden reference HV, the 3W resistors between the ksa1156es and ksc5026 run close to 90C and dont have a lot of room around them, on one side there is a large 0.1uF 1KV film cap and on the other 5 closely packed transistors...

So the T2 needs good heat sinking and the psu boards could do with lengthening a little to give the 3W resistors a little more air flow around them and drill holes under them.


Update DC offset fix,

The 5K pot in parallel with the 510 ohm emitter resistor gives a minimum resistance of about 461ohm which still gaves me  ~ -33V offset. With the 5K pot in the middle you get around 423ohm combined resistance and ~ -85V offset. So I replaced the 510ohm with 604ohm and this gives enough adjustment range to get <1V DC offset.  (I would have implemented joamats suggestion but I did not have any 10K pots)

Edited by jamesmking
  • Like 4
Link to comment
Share on other sites

On 9/9/2020 at 9:11 AM, jamesmking said:

I found my first issue, the DC balance between  + and - halves of a channel is fine <50mv BUT the DC offset to ground started at -85V and even adjusting the offset pot all the way I only got it down to -33V.

My experience is that the balance servo is very efficient so 50 mV is expected. As for offset – 5K trimmer parallel with 510R resistor reduces highest emitter resistance to roughly 465 ohm which probably is too low. May I suggest replacing R73 (6.2K) with a 10K trimmer and keep R64 (510R) untouched? I have used that for a long time and it works great.

Edited by JoaMat
  • Thanks 1
Link to comment
Share on other sites

T2 first listening impressions.

wow..... erm.... wow..... erm..... wow

The first thing that strikes me is marco dynamics are amazing, on recordings that are not that dynamic the amp does not exaggerate the dynamics, but on dynamic recordings like full orchestra the dynamics are so powerful but not overdone.

This is backed up with bass is far more defined, less mushy, more full than my other stax amps but the bass is not not dry or hard or overblown

Stereo images are more solid and tangible and more precise than any other stax amp I have, its so close to the experience my Quad 2805 speakers.....

Detail is amazing, but the detail is not thrown at you and its therefore not tiring to listen to. But I do notice tape edits, dropped violin bows, musical score page turns more... but it just adds to the experience that tens of musicians in a room can't be totally silent all the time.

when music gets complicated there is far less tendency for sounds to just clump together into a messy mushy ball. However you still get the feeling of musicians playing together and the detail is not analytic. The amp just says here is all the detail let it wash over you and you decide how much you want to listen into the detail and how much you want the whole music.

This is the first time I have been satisfied with the bass and dynamics from my sr007a mk2. I think it shows that my stax sr007a just need as much drive as they cant get.

My journey: stax srm006 (piece of mushy yet bass light and less bass the higher the volume, undynamic, undetailed, poorly constructed, stereo thick soup, gloop with the sweetness of a lemon), hi-amp alpha centauri (improvement in every way over the srm06, its an upgrade but you still feel there is music missing. Overall you get the impression is nothing special and has but poor bass, slightly deeper than the srm006, but mid and upper bass is almost non existent. Bass loss with volume is far less), blue hawaii (murders the alpha in every way but maybe slightly hard sounding for me and a bit of a tiring long term listen. It was a very enjoyable build and a considerable step up from the alpha but I rarely could relax into the music), Mini T2 (better micro dynamics, smoother and more refined more listenable than the hawaii but lacking in some drive and macro dynamics, detail and bass. I still got the feeling with it that it was struggling with the sr007s a bit) and finally the T2 (im sure I will eventually find something wrong with it... that's just the way my brain works but for now im content [which is rare])...

So whats next after the T2? (other than building a second one)....

Many many thanks to Kevin for the T2 and joamat for the modified T2 and mini T2, spritzer for providing the nos transistors and everyone for their encouragement and help.



Edited by jamesmking
  • Like 9
  • Thanks 1
Link to comment
Share on other sites

  • 2 weeks later...
41 minutes ago, JoaMat said:

Nice power supply. I can’t read the labels on trafos. What are their specifications?

low voltage transformer 52mm high 105mm diameter ( I like to run the heaters at a bit under 6.3V AND in my experience the transformers I buy provide a little more voltage than the specs)

2x 4.5A @ 6V for el34s

2x 1.5A @ 6V for 6922s

2x 0.4A @ 15V for balance servo, HV delay circuit and delay relay

high voltage 65mm high 125mm diameter (voltages are the same as the psu board silkscreen.)

2x 0.15A @ 260V for +250, -260VDC

2x 0.15A @ 450V for +-500VDC

1x 0.15A @ 465V for -560VDC

Both transformers have magnetic shields, electrostatic shields and thermal resettable fuses. transformer manufacturer Mueller rondo in germany http://www.mueller-rondo.com/kontakt.htm

With postage and vat slightly under 200 euros for the pair.

I like seperate HV and LV transformers so I can run the heaters and LV at full voltage and slowly bring up the HV on a variac when I am doing initial testing. It also offers the chance to easily reuse the LV transformers in other valve amps.

Edited by jamesmking
  • Like 1
  • Thanks 1
Link to comment
Share on other sites

  • 3 months later...

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


  • Create New...