[vintageaxeman]

I do the same )

I just published a modification for the Stax SRM-T1/T1S/006 in AudioXpress July 2017, which is out now. The modification adds 5.1 kilohm output safety resistors and cascoded constant current output loads, which with their heatsink fits the space vacated by the plate resistors. It uses the stock 6CG7/6FQ7 output tubes.


Unlike plate resistors, constant current loads burn up practically no signal current, so while the stock amp burns up 9.2 mA signal current in the plate and feedback resistors at its specified 300 VRMS maximum output, the modified amplifier only uses 2.8 mA driving the feedback resistors, leaving much more current available to drive the headphones. Since the output tubes are much less stressed, distortion is decreased and 2 dB of extra headroom gained. This modification can drive even the demanding SR-007, something the stock design strains to do, and improves bass punch, overall clarity and neutrality. Think of it as a T1 on steroids.


The modified circuit is very similar to a KGST. This is no accident, as the KGST was designed to be “a Stax SRM-007t with no cost or retail considerations,” and the 007t is a higher power T1.


The modification lacks the KGST’s regulated power supply, but larger power supply caps have been fitted. Since the amplifier is fully differential pure class A with current sources or loads at every stage, the current draw is pretty constant, minimizing power supply effects.


Replacing all the old electrolytic power supply caps, as outlined earlier in this thread, should also be done as routine maintenance. Parts cost for the modification is about $35, so total parts cost is around $100. With T1 amplifiers going for $400-$700, this is the best bang for the buck, a good starter project for someone wanting to get into electrostatic amp DIY without going to a full build.

Is there a similar mod for a 700Tii please?

 

[JimL11]

Not specifically, but the idea is the same. Are you referring to the 007Tii (uses 4 6FQ7 tubes) or the 700T (uses 2 6SN7 tubes)? Both have essentially the same circuit as the T1 except for the output tubes, with the same input and intermediate stages. The amount of standing current in the output stage varies between the amps. Note that the T1 and 007 use the same output tube, and that the 6FQ7 is basically a nine pin miniature version of the 6SN7 with practically identical tube parameters. The 007Tii uses double the number of output tubes compared to the T1, but with the output plate resistor reduced to 47 kilohm from 66 kilohm in the T1, so the output stage standing current is increased about 40%. I don't know what the plate resistor value is for the 700T. In any case, you can calculate the approximate value of the constant current load current by taking the power supply voltage, which is about 350V, and dividing by the value of the plate resistor. For example, for the T1 (which has a PS voltage of about 320v), the current should be 320/66 ~ 4.9 mA, for the 007Tii it is 350/47 ~ 7.4 mA, and so on. Given that the circuits are basically the same, my feeling is that the best value amp is the T1/006 series, and that the new 700T is overpriced, but of course that is only my opinion.

In either case, the idea is to replace the plate resistors with constant current loads mounted on a heatsink. You just have to calculate the correct standing current, and figure out where to mount the heatsink for the current loads. This mod more than doubles the amount of signal current available to drive the headphones as you are no longer wasting it as heat in the output plate resistors, but of course it voids the warranty. It really is a pity that Stax haven't instituted this mod themselves - it's a much bigger improvement than the 700T is over the 007Tii. It's not that they don't know how - they used constant current loads in their T2 and in all their current solid state output amps.

 

[vintageaxeman]

Thank you for that wonderfully detailed reply. Mine is the 007Tii. It’s a secondhand unit so it’s not under warranty. What are my options for getting this mod done?

 

[JimL11]

Are you looking to do it yourself or are you looking for someone to do it for you? Where are you located?

 

[vintageaxeman]

I think I would need a skilled person to do it for me. I'm ok with soldering simple circuits like crossovers, but I probably wouldn't attempt this. I'm in the UK.

 

[JimL11]

Well, it's not that complicated a circuit - 2 MOSFET's and 3 resistors plus a trim pot per current load - you need 4 current loads, which replace the plate resistors. I hard wired everything and attached it to a large single heatsink, but if that's too complicated, Kevin Gilmore has free circuit board plans (which you can send to any circuit board manufacturer) that includes space for heatsinks for each current load. You have to figure out where to mount the circuit board - probably to the side of the amp near where the replaced resistors are.

OTOH, if you want someone to do it for you, I believe that @Firschi can do it - he's a Head-Fi member in Germany. Or else he might be able to direct you to someone who can. He also has a website which discusses how to do the mod in detail. With the 007, the 5.1k output safety resistors and the 5 megohm bias resistor are already in place so you don't have to bother with that part of the mod.

Hope this helps.

 

[vintageaxeman]

It certainly does help. Thank you. I’ll follow it up! Cheers.

 

[kevin gilmore]

stax 400s, 500t and 700t all would benefit with the current source board. does not look like the current board will fit any of these, best to
make a pair of boards each with 2 channels on them. The mounting holes would need to be carefully measured for me to make new boards.
each of these would need different boards. 700t seems to be the easiest.

 

[Zoide]

Would a CCS modded T1/T1S (with or without ECC99 tubes) be better than a 323S / 353X?

What about a modded 717 / 727?

Thanks

 

[JimL11]

Would a CCS modded T1/T1S (with or without ECC99 tubes) be better than a 323S / 353X?

What about a modded 717 / 727?

Thanks

That is a subjective question so it depends a bit on what your sonic preferences are. I had a 727 with global feedback mod, and I preferred my modded T1, which sounded a bit smoother. I sold the 727, still have the T1. A couple friends who heard both also preferred the T1. The modded 727 does have a bit more power and current capability, is supposed to be a bit better than the 717 due to using current sources in a couple places where the 717 used resistors. The 717 and modded 727 are supposed to be better sounding than the 323S or 353X. Of course with the solid state amps, don't have to worry about leaving on all the time, or eventually needing to replace the tubes. Tube life is good with the Stax amps as they are run very conservatively.

Haven't heard a T1 with ECC99 tubes, which require rewiring the tube sockets, so have no opinion on the benefits or lack thereof compared with the original 6CG7 tubes of the T1/T1S. If you go for the ECC99 mod, you can bump up the standing current about 10% based on the 600LE amp, which I believe is the only Stax amp that used ECC99 tubes, and IIRC used 60k plate resistors in place of the 66k resistors in the T1. I have seen a couple of comments that the 600LE sounded better than the T1 series.

 

[Zoide]

That is a subjective question so it depends a bit on what your sonic preferences are. I had a 727 with global feedback mod, and I preferred my modded T1, which sounded a bit smoother. I sold the 727, still have the T1. A couple friends who heard both also preferred the T1. The modded 727 does have a bit more power and current capability, is supposed to be a bit better than the 717 due to using current sources in a couple places where the 717 used resistors. The 717 and modded 727 are supposed to be better sounding than the 323S or 353X. Of course with the solid state amps, don't have to worry about leaving on all the time, or eventually needing to replace the tubes. Tube life is good with the Stax amps as they are run very conservatively.

Haven't heard a T1 with ECC99 tubes, which require rewiring the tube sockets, so have no opinion on the benefits or lack thereof compared with the original 6CG7 tubes of the T1/T1S. If you go for the ECC99 mod, you can bump up the standing current about 10% based on the 600LE amp, which I believe is the only Stax amp that used ECC99 tubes, and IIRC used 60k plate resistors in place of the 66k resistors in the T1. I have seen a couple of comments that the 600LE sounded better than the T1 series.

Thanks, I'm just wondering about the best next step up from my SRM-323S. I'd like to get an amp with balanced inputs and enough power to drive the Nectar Hive X, which I'm thinking of buying to complement (or even replace) my L700.

 

[heatsink]

First of, I'd like to thank all the folks participating in this awesome community from all the different countries, JimL, Spritzer, Kevin Gilmore, Sören, Firschi, you name them. While being very new to head-fi, I'm not entirely new to tinkering with STAX devices.

Back in late 2020 I replaced all the electrolytic capacitors of my SRM-006t and added CCS to it. Later I went for ECC99 tubes, following the advices given in this thread and the german blog that somebody posted a link to some pages back. Everything went well and sounded brilliant with my lambdas.

This year I traded my modded 006t for a used SRM-007t, adjusted the input voltage and replaced all electrolytic capacitors. A few months later I successfully performed the rather comprehensive work for it to take 6S4A tubes. Based on my experience with CCS, I decided to add this feature to my 007t as well. That's where I'm struggling currently.

I'm using the same schematic for the CCS that I've been using for my SRM-006t. This get's closest to what I use: http://blog.prof-x.de/wp-content/uploads/2018/08/20180804_CSS_JimL1.jpg I replaced the unlabeled resistor by a 500R trimpot.


Of course, this time I adjusted the outputs to 8.0mA each. For some reason I keep frying the fets in my CCS. Parts were purchased from Mouser, I would consider them genuine. First I though this was a temperature issue, but even when using massive heat sinks on the 10M90S, both fets (10M90S and DN2540) keep failing, from time to time. By failing I mean, that one or both fets in each cascade seem to lose the ability to restrict the current flow, leading to way too high CCS output currents up to 20mA (instead of 8 what each CCS is being adjusted to). Of course, the power supply can't provide that amount of current without B+ dropping. My current working hypothesis is that the gates of one or both fets of each CCS get damaged, either during power-up of the amp or after switching power off. 10M90S are hard to find, and I'm in urgent need of a solution for this so that I don't keep frying them.

Does anybody have an idea about how to fix this? Any help is highly appreciated.

 

[kevin gilmore]

you are using the to220 version of the dn2540?

 

[heatsink]

I've been using the TO-92 version exclusively. Haven't had any issues with that in my former SRM-006t so far.

 

[JimL11]

Mouser is a reliable supplier - that's where I get my FETs.

So there are a couple possibilities. The DN2540 regulates the current, so if the current is much higher than set, it likely means the DN2540 is blown. In theory the TO92 version should be enough, since it is the bottom FET, so it should only see about 3 volts as long as it is sitting between the grid and source of an intact, working 10M90S, and with 8mA, that only amounts to 24 mW, which is way below its power rating of 1 watt. So even the TO92 version of the DN2540 should be fine - as long as the 10M90S is intact.

However, if the 10M90S goes south, the DN2540 will fry as well. The power supply voltage of the 007 is +/-350V which shouldn't be enough to blow the 10M90S, which is rated for 900V, so the only way it can be damaged is if its power capacity is exceeded, which makes me wonder how "massive" your heat sinks are. The CCS heatsinks in the 007 need to be about 60-70% bigger than in the 006 because they are dissipating that much more power (8 mA ve 4.8 mA using 6CG7). If you're using the same heatsinks in the 007 as in the 006 that could be the issue.