DIY mini T2 Build Thread

[mwl168]

I am starting this thread for all that are or will be building JoaMat's mini T2 to share information, ideas, thoughts and experiences.

The attached schematic is an updated one .The resistors in the BOM are thick film ones and can be replaced with thin film ones for likely better outcome.

The power supply requirements for this amp are +400VDC, -460VDC (EL34 output stage), +220VDC (6922/6DJ8 front end), +15VDC, -15VDC (servo, cathode current source) and 6.3V filament supplies for the 6922/6DJ8 and EL34 tubes that can either be AC or DC. I also plan to use the +15VDC for the CPC1117N high tension delay.  

EDIT (Feb. 2020): updated schematic and BOM (Thanks JoaMat!)

 

Note: PCBs pictured below are of older version.

miniT2 v22 schematic.PDF

miniT2 v22 BOM -Date_2020-02-16.xlsx

Edited March 18, 2020 by mwl168

[Inu]

I am planing to use Toshiba TTC004B instead of KSC2690A.

Cob is lower than KSC2690A  (12PF - 19PF) and better hfe linearity.... and I prefer Toshiba rather than KSC/KSA .  

Edited February 24, 2019 by Inu

[mwl168]

Forgot to mention that it’s recommended that the +220VDC be supplied by its own regulator instead of using a drop resistor from the B+. The top device, HN4A51J, is a dual transistor chip and need to be careful not to overheat it.

Edited February 23, 2019 by mwl168

[chinsettawong]

Interesting.  How is the sound as compared to the full size T2?  

[JoaMat]

I don’t really want to make any comments on how things sound and compare them. Mainly because it makes me feel uncomfortable and I also lack the ability to analyze the sound. Anyhow I’m very happy with “mini T2” and I want to think that it is quite similar to THE DIY T2. Now I have put the original DIY T2 in the closet and I don’t miss it. I still have my modified T2 in our summer cottage and I like that one as much as the “mini T2”.

But I have three more original T2 boards and all sands for them – no duck feeding the next couple of years.

[Mirko]

how big are the size of the boards? Looks like there are some space that can be further shaved off...

[JoaMat]

Board size 160mm x 100mm. Sure it's possible to shrink it, but that will be the micro T2, then we have nano, pico …..

Thinking about the possibility to use an unmodified T2 PSU.

If applying +250V instead of +220V you get +79V at upper 6922’s grid and +230V at anode (with no changes on the board). T2 has +69V resp. +200V. Power at HN4A51J goes up from 2 x 76mW to 2 x 87mW, so still OK I think. Question is if the higher tube voltages make any difference. But I don’t think it’s wise to put 48V cross HN4A51J when bringing down anode voltage to +200V (with +250V supply).

As for +/-500V; no problem with 01N100D it’s a 1000V device.

I do think it might be possible to use original DIY T2 PSU. Has to be proven.

Edited February 17, 2020 by JoaMat

[chinsettawong]

Could you kindly share the Gerber files?  Or are there anymore boards to be grabbed?

[Laowei]

11 hours ago, Inu said:

I am planing to use Toshiba TTC004B instead of KSC2690A.

 

Any advantages to do this?

[mwl168]

More information on the PSU re: current draw on rails;

for the +220VDC, plan on around 20 - 30 mA draw

for +400VDC and -460VDC rails, plan on about 60 mA draw

These are two channels combined based on information from JoaMat. Need to add headroom on top. 

[JoaMat]

1 hour ago, chinsettawong said:

Could you kindly share the Gerber files?  Or are there anymore boards to be grabbed?

I want to make some minor changes to the layouts. After that I think more boards will be available here. Should not take long.

[Laowei]

Please consider labeling the title and actual file with the running revision level on: 

1) schematic.PDF

2) BOM.xlsx

3) Gerber and PCBs. 

Otherwise this could quickly get messy....

Edited February 23, 2019 by Laowei

[mwl168]

I have no more boards available but have JoaMat’s blessing to run another round of GB if there in enough interest. 

Please raise your hands here if you are interested.

Keep in mind the mini T2 are mirrored left and right channel boards so it’ll take twice the number to reach the same economic scale for the volume discount on board cost.

[JoaMat]

OK – Connected an unmodified T2 PSU. Raised voltages to +250 and +/-500V. Top 6922 anode voltage sits at 227V and grid at 78V. So far it works. Will run this for a while and check for darkening spots on the boards.

[Laowei]

Received my mini T2 and CFA PCBs from mwl168 this morning. These are excellent quality!

Many thanks to Michael and Joachim! 

[blubliss]

I'd take some boards.

[Rossliew]

Is the cost of building the mini T2 (much/slightly) lesser than a full sized one? 

[JoaMat]

A few points regarding the boards Michael have had made on my gerbers. They are basically the same as my kitchen made boards I made  five months ago.

• The small tube footprint has no hole for the center pin. So far I’ve cut away the center pin on the small tube socket after checking the pin holders to be not to tight nor to loose. One time I got tube sockets where the tubes fell out when turning upside down. So either you have to cut pin or drill a hole. Be aware that upper tube has a trace between the grids that might be to close to center pin.
• Right board 270K resistor has a 0805 pad (intention was 1206). It’s possible to solder a 1206 resistor on the 0805 pad. If +220V you get about 130 V cross the resistor so Vishay CRCW0805 is OK here. But with +250 V you get close to CRCW0805’s limit of 150 V.
• Some of component values on silk screen are poorly place and almost unreadable. The blame is entirely on me, I’m very sorry.
• There are four octagons on bottom side. One at each big tube socket center hole and one each side of small tubes filaments connections. Here I have standoffs.

A few comments on updated layouts.

• Small tube footprint with center hole.
• 270K resistor pad corrected - now 1206.
• Silk screen remade. Pleas notify me if there is any left that needs additional attention.
• Now “value” of connections on both sides (except heaters). No terminal blocks on my boards.

Top right. Updated 2019-02-27.

Bottom right. Updated 2019-02-27.

Top left . Updated 2019-02-27.

Bottom left. Updated 2019-02-27.

I call above version .2. My kitchen made and Michaels are version .1.

Regarding this amplifier - there is nothing I've invented. I just have picked up different things and tried to put them together. As:

• Opto servo (the only offset servo you need) by Kevin.
• 01N100D/TL1021 current source by Kerry.
• Schematic of Stax T2 provided by Kevin, Et. al.
• Schematic of Stax T8000 provided by Kevin and Birgir.
• Etc. etc.

Thanks for all knowledge, information and inspiration from you DIYers and Head-Case members.

Please look for errors or other issues.

Edited February 27, 2019 by JoaMat

[JoaMat]

R9 is 255R on silk screen on the boards now sent out by Michael. For some time I have run the amplifier with 200 ohms which increases current through the small tubes a bit. I haven’t made any measurements on the amplifier to verify currents but 200 ohms will give something like 3 mA through tubes (T2 has 5 mA). I do think the change improves the amplifier a bit.

As I want voltage at base Q8 and Q9 to be 20 volts above -HV, same as the real DIY T2, I’ve today increased R13 and R14 from 200K to 330K. I don’t get +20 volts but it’s closer to it now.

R9 on latest schematic is 200 and I think it should be so.

R13/14 are 200K on schematic. To follow the DIY T2 you may increase it to 330K. But maybe it doesn’t matter which you choose. Any thoughts out there?

[JoaMat]

I wanted to find out the currents “produced” by Q1 and Q2. Since I don’t have the courage to stick my nose or test pins in a live amplifier I made this test board today.

Measurement on the test board indicates voltage cross emitter resistors at R1, R3 and R9 close to 1.10 V. If  we actually have 1.10 V than current through small tubes are 2.6 mA and not 3mA that I mentioned in previous post. The Q1 current sources give 3.8 mA each source.

I’m considering changing R1 and R3 to 270R and R9 to 180R.

Now days I check for the polarity of the LED TLMS1100 by using DMM in diode testing mode. Positive pin on anode and the other on cathode you get a red light. But be careful the LED will easily pop away and it’s not sure you ever will find it again.

[G600]

Excellent work as usual, JoaMat.

Please let us know if you test those values in your working unit !

I'm thinking of populating the SMD resistor with the fine Susumu RG (thin film). Used them in the QRV-08 (Sjostrom).

1206 are 200 V rated and 805 150V, should be plenty.

[JoaMat]

Rushed away to my local supplier and arrived just in time before they closed for the day. Got 0805 180R and 270R. At home they turned out to be 0603 but I managed to get them in place. On the test board I get  exactly what I expected and I assume that that’s the case  with the amplifier as well. Measured 18.5 V  above -HV at bases of Q8 and Q9  with R13/14 @ 330K.

I think Susumu RG is a good choice. ToshibaTTC004B is interesting – thanks INU. The smd transistors are also Toshiba.

[G600]

Do you see critical resistor, some that will benefit from lower ppm or tolerance ?

 

[JoaMat]

No I don’t. But that’s mostly because it’s beyond my understanding. Maybe someone out there can give us a hint.

Made some rework regarding resistors value. Not much to say about it. The R24 and R25 - you probably won’t find 800R easily. Now changed to 787R. The resistor determines the current of CCS. Divide LT1021’s 10 V  by resistance and add 1 mA and you get the current. 787 ohm give 14 mA.

The four pictures (Top right/left and Bottom right/left) a few posts up are now updated to reflect new values.

Schematic for the latest board layouts.
Schematic mini T2 version .2.pdf

Edited February 27, 2019 by JoaMat

[JoaMat]

On 2/23/2019 at 6:33 PM, mwl168 said:

Keep in mind the mini T2 are mirrored left and right channel boards so it’ll take twice the number to reach the same economic scale for the volume discount on board cost.

The two mirrored boards are now a married pair. Size 6.3in x 7.87in / 160mm x 200mm. If you want  them divorced use a guillotine and cut along the center line.

Edited February 28, 2019 by JoaMat