SMT-334 Temperature Compensated for Better Image Stability

The LM334 as implemented in amps using it for phase splitter bias suffers from getting too hot and thermal drift.

Temperature stability can be improved by mounting under the board and made near perfect by compensation.

Miller Audio has introduced a DROP IN SMT version called the SMT-334 which solves phase splitter drift.

SMT-334






I have ZERO affiliation with Miller Audio. I found the link here.

I was too looking into this issue.  I came across this site: https://www.audioamp.eu/the-problem-boards-vta-st-70-with-lm334-is-solved-ccs-zero-temperature-coefficient-current-source-for-tube-phase-inverter-en/?nastav_menu=USD.
Back in Jan 17-2024 I copied the page and the cost was just over $5.00, now it's $9.74. Talk about inflation!
Also at that time they were out of stock, so I decided to do a PWB using thru-hole parts that I had on hand.
   

I decided to use LTSpice and model the performance of the 334 with and without compensation. Here are the results:
Without compensation:


As seen from the results the change in ma from 10C to 100C is about 7.1ma to 9.4ma, a delta of 2.3ma. The 9.4 ma puts it close to the spec limit of the 334. With tolerances figured in it is possible
to exceed the limits of the 334 and may cause a problem, although in actual practice there has been little issues.

With compensation:



As seen from the results the change in ma from 10C to 100C is about 7.4ma to 7.7ma, a delta of 0.3ma. The 334 now stays well within it's rating.

Notice any listening improvement?

What amp do you have?

Seamus wrote:Notice any listening improvement?

What amp do you have?

Quite frankly, I have not noticed any difference in sound, I am not blessed with hearing acuity or "Golden Ears" as they say to pick up on any differences.
This being a hobby of mine, I tend to fix things that are not really broken

I have the MK3 amps

jwb474 wrote:This being a hobby of mine, I tend to fix things that are not really broken

HaHaHa!!!

jwb474 wrote:I have the MK3 amps

Location and current could play in audibility. According to page I referenced, the case temperature fluctuated considerably
Where is the LM334 situated on the driver PCB and original current?

" in Jan 17-2024 I copied the page and the cost was just over $5.00, now it's $9.74. Talk about inflation! "

That's not inflation; it's increased demand.

I did this modification in the following way. Mounted LM334 on bottom of VTA driver board (M125). I bonded the diode with silicon glue to the. LM334 and used a few precision resistors, using VTA driver board.

sonically, I like my amps much more since I did this change, although it would be difficult to pin it on this change, as I added a bunch of stuff. IMHO worth doing to improve stability of the phase splitter on VTA board. IMHO, using a component like lm334, which is known to have temp instability, it seems a no brainer. I doubt the OG driver boards use any sand....

Rec wrote:I did this modification in the following way. Mounted LM334 on bottom of VTA driver board (M125). I bonded the diode with silicon glue to the. LM334 and used a few precision resistors, using VTA driver board.

That's the best way to get the best compensation. My layout, referenced above, would allow me to stand the body of the diode up next to the LM334 and use thermal epoxy to secure the diode body to the LM334 body.
Maybe someday when I don't have as many irons in the fire and I get industrious I will do that and figure out a easy way to measure current vs temp in actual application

jwb474 wrote:... figure out a easy way to measure current vs temp in actual application

Measure the no signal voltage drop across the two plate resistors after the amp temperature has stabilized.

Here's what I did FWIW. The resistors fit nicely on the top side, and the diode was mounted on the bottom side.

maltboy1 wrote:Here's what I did FWIW. The resistors fit nicely on the top side, and the diode was mounted on the bottom side.

That's a great looking mod.

maltboy1 wrote:Here's what I did FWIW. The resistors fit nicely on the top side, and the diode was mounted on the bottom side.

The diode should be in as close thermal contact as possible to the 334 so the change in the diode vF with temperature compensates for the same property in the 334. Ideally with a thermal compound so as to make the two components one.

jwb474 wrote:That's a great looking mod.

A huge problem in HiFi today is looks over good engineering.

The M-125s are anything but elegant but when properly built, OUTSTANDING!

Seamus wrote: The diode should be in as close thermal contact as possible to the 334 so the change in the diode vF with temperature compensates for the same property in the 334. Ideally with a thermal compound so as to make the two components one.

That would be best. Looks like it would be easy enough to stand up the diode and lead form it so it would contact the 334.
This is the thermal epoxy I use: https://www.digikey.com/en/products/detail/mg-chemicals/9460TC-3ML/13997491?utm_adgroup=&utm_source=google&utm_medium=cpc&utm_campaign=PMax%20Shopping_Product_Medium%20ROAS%20Categories&utm_term=&utm_content=&utm_id=go_cmp-20223376311_adg-_ad-__dev-c_ext-_prd-13997491_sig-Cj0KCQiA88a5BhDPARIsAFj595gqN7piVoLAF-gTipAjsgbggzECvnf3pXR-MIKSL8HphDqfK9-g5xwaAs5cEALw_wcB&gad_source=1&gclid=Cj0KCQiA88a5BhDPARIsAFj595gqN7piVoLAF-gTipAjsgbggzECvnf3pXR-MIKSL8HphDqfK9-g5xwaAs5cEALw_wcB

Here's a link to MG Chemicals info page on thermal adhesive page: https://mgchemicals.com/category/adhesives/thermally-conductive-adhesives/

jwb474 wrote: A huge problem in HiFi today is looks over good engineering.

This is so true. Back in my engineering career days (A long time ago) the wishes of the marketing department often trumped the engineering department as long as the unit under design met their specs.

Received LM-334 boards from Pavel and installed them this morning in the ST-120.  Had to mount them top side as the large stepped attenuator used restricts bottom side mounting. The amp still takes an hour or so to warm up to sound best but do believe I have slightly firmer bottom.  Anyway should have better stability over a wide temperature range versus stock.