hamptone burning resistors...

[pucho812]

So I bread boarded some of the hamptone fet sections and have been trying to use them with a bi polar supply.  The bi polar supply I am using to test these with does +/-17vdc I've subbed r6 as I only had a 6.8Mohm and r5 as I only had 143 ohm on hand. R5 is still a half watt though. I also subbed the bd139 for a zxt651 as again all had on hand.  Anyway when connected to the testing PSU I keep burning out r5. it smokes and dies. Everything else in circuit tests out  as working and is wired per the above schematic.  I have yet to run audio through it as well r5 keeps smoking and burning out.  Any ideas?
I
This was originally noted  in this thread here
http://groupdiy.com/index.php?topic=40106.msg496142#msg496142

And for short cut
here is the modified schematic

 

[JohnRoberts]

Looks like maybe one or more parts are missing?

I'd add a resistor from base of Q3 to ground or -12V

JR

 

[12afael]

if the center of the output trancistor pair is at 17V then at the emitter resistor you will have 17/143 =118mA then 118mA at 17V it is around 2W. the bias point is set with the 2kohm pot at the jfet stage. if you set that stage for maximum voltage excurtion R5 should be between 1 and 3V I guest.

 

[12afael]

wait I did the test with 0 and 34V , r5 should be then from -16 to -19V

 

[PRR]

> hamptone fet sections and have been trying to use them with a bi polar supply.

For gosh sakes, WHY? It is a single-ended circuit. Using bipolar is complicated, and injects negative-rail crap directly into the input.

Use just the zero and +17V. It'll work.

> I keep burning out r5

The original circuit is quite specific. It isn't designed to be built with odd parts and odd voltages.

You have increased the supply from 24V to effectively 34V. That *doubles* the power dissipation.

You have changed Q3. Q3's bias is *super* sensitive to Q3 hFE. I can not find an hFE spec for the original ZXT651. The BD139 comes in three hFE dash-codes; I dunno which you used and you probably don't either.

*Ass*u*ming* Q3 hFE is 100, and 34V supply, R4 passes 0.7mA into Q3 Base, which is multiplied by hFE=100 to make Q3 pass 70mA. 70mA in the original 100r makes 0.49 Watts. You changed to 143 Ohms, so 0.7 Watts. If Q3 hFE is 200 (it can be 250), then 140mA and 1.4 Watts.

Seriously: put it at single 17V supply until you are ready to do it just as Scott documented.

 

[PRR]

> maybe one or more parts are missing?

That's the way Scott wrote it up.

http://www.diyrecordingequipment.com/wp-content/uploads/2011/02/hamptone_fet_scan.jpg

It is hand-tweaked, not production-robust (and not stray-parts robust).

 

[usekgb]

The Hfe of Q3 needs to be very low, like 100-150, otherwise there will be too much current flowing through R5.  Try adding a 50k pot in series with R4 to trim the output current.  This should be roughly 40mA across R5.  With a bi-polar supply, this thing won't be running single ended class A anymore.  You may need to use a FET that can handle higher voltages, like the 2N4393.  The bias point of the FET will likely not be at 12V anymore, due to the change in power supply.

I think PRR is right on this one.  Use a single ended +24V supply, and be done with it.  I just finished one up a couple days ago, and it is awesome!  I subbed a few parts for what I had on hand.  I used a 2N4393 fet, MPS-A12, and a different Q3 with an Hfe around 150.

 

[JohnRoberts]

> maybe one or more parts are missing?

That's the way Scott wrote it up.

http://www.diyrecordingequipment.com/wp-content/uploads/2011/02/hamptone_fet_scan.jpg

It is hand-tweaked, not production-robust (and not stray-parts robust).

Sorry it's against my design ethos to allow such high sensitivity to device parameters.  I've been bit too many times and seen others go down in flames..

JR

 

[usekgb]

Yeah.  This thing is really tweaky.  Definitely not for production runs.  In fact, it performed horribly on a breadboard.  It wasn't until I cased it up and star grounded everything that it actually worked correctly.

 

[pucho812]

errr  when I did sim on the circuit it did worked which is why I did the bread board version. The idea was to have a 500 series version of the hamptone. As PRR suggested, run it at 24VDC only. 

Since the API racks run +/-16VDC

Would something like this work to get me my24VDC?

 

[Dylan W]

The "best-practice" approach for 500 series +24V (EG the way the most of the Neve-alikes do it) seems to be using 7908 on the -16V rail to get +16/-8. There are a few threads about this.

 

[Dylan W]

Good info here.

http://groupdiy.com/index.php?topic=54891.0

 

[usekgb]

Another thing you may want to do, to allow variations in the FET, is to place a 2k trimmer in series with a 1.8k resistor, all in place of R3.  This will allow you to dial in the bias point of the FET.  With just the 2k trimmer, you may not be able to get the bias point down to 12v.  For Q3, I used an MPS6602, as I had a bag of them laying around.  For R5, I used a 1W 100R resistor.  During initial testing, this resistor got quite hot.  I believe this was due to the fact that my circuit was oscillating badly.  Once I got everything stable, R5 was cool to the touch.

 

[PRR]

> Would something like this work to get me my24VDC?

Draw-out your "0V" ("ground"). It does not go where ground for everything else in the box goes.

> the way the most of the Neve-alikes do it

The Nevealikes are *transformer*-coupled in and out. The audio interfaces are not referenced to the power supply.

Run the Hamp on +16V. It will work with small changes. Accept that it will not deliver +15dBm like the 24V version can.

> too much current flowing through R5

I would even go 220K (for 34V) or 150K (for 24V) to cover high-hFE Q3. Look for 4V across a *100 Ohm* R5, and gently go-down on Q3's base resistor.

 

[pucho812]

did a replace on components with the correct ones, correct values. No more smoking resistors.  It now runs off my bi-polar test supply. time for some audio testing...

 

[pucho812]

Interesting ???  I can run each of the two fet sections by themselves and they sounds fine with 1K. So for example going in unbalanced, through either module out to a pot then to an output and it sounds fine. O-scope shows some junk  but since it's a bread board sitting on the bench i am not surprised. The moment I go from the 10kb pot into the second module, all hell breaks loose. I have a small range in the pot throw where I get nice volume the the rest before and and after is mostly noise and all kinds of garbage.  hmmmmmmm...

Any suggestions?

 

[usekgb]

Mine did the same thing on breadboard.  Did you include the 33ohm resistor and 470uF pot going to each gain stage?  This helped mine.  It still took off until I put it in an enclosure though.

 

[PRR]

> runs off my bi-polar test supply. ...I have a small range in the pot throw where I get nice volume the the rest before and and after is mostly noise and all kinds of garbage.  hmmmmmmm...

Hmmm indeed. You ask for advice, you ignore it. 

 

[pucho812]

Prr, your advice is always appreciated , it's such a learning experience and is my goto when I have this fail in a day or two trust me

 

[usekgb]

With this thing running on a bipolar supply, are you biasing the fet at 0V?  How are you avoiding crossover distortion?