Test startup issues.
I have two pcb's fully populated and connected to my test power supply. When I first started applying the voltage with the Variac R-13,R-65 vaporized. I turned everything off and it seems no further damage is noticeable. I only tested one pcb at this time.
I applied the voltage again to full spec. There was no more magic smoke. When I connect up the speaker and input I have quite a bit of 600-1000 HZ. hum (using my Fluke meter). At this time I am going to attribute that to all of the flourescent lights, fans and improperly twisted and routed test wiring.
I shut everything down so as to better design the test layout and rule out this as a cause for the hum. I also wanted to see if anyone had a solution to the 2 vaporized resistors which are supposed to be part of the limiter. They were 1/2 watt 10R metal film items.
I really hope there is not a big assembly error. I am overly cautious in this area since my troubleshooting skills are lacking. I have triple checked the orientation of all of my discrete components. Checked for any pieces which my contact ground or each other. The closest I came was 7 meg ohms. I do not think this should be a problem.
Anyone have any suggestions???
Tad
I have two pcb's fully populated and connected to my test power supply. When I first started applying the voltage with the Variac R-13,R-65 vaporized. I turned everything off and it seems no further damage is noticeable. I only tested one pcb at this time.
I applied the voltage again to full spec. There was no more magic smoke. When I connect up the speaker and input I have quite a bit of 600-1000 HZ. hum (using my Fluke meter). At this time I am going to attribute that to all of the flourescent lights, fans and improperly twisted and routed test wiring.
I shut everything down so as to better design the test layout and rule out this as a cause for the hum. I also wanted to see if anyone had a solution to the 2 vaporized resistors which are supposed to be part of the limiter. They were 1/2 watt 10R metal film items.
I really hope there is not a big assembly error. I am overly cautious in this area since my troubleshooting skills are lacking. I have triple checked the orientation of all of my discrete components. Checked for any pieces which my contact ground or each other. The closest I came was 7 meg ohms. I do not think this should be a problem.
Anyone have any suggestions???
Tad
Andrew,
I am already using the bulb tester. I have not turned the bias down as you suggest. I will most certainly start with the bias adjusted much lower. I had forgotten that the variable resistors do not always arrive at a predetermined value. When I went back and checked I was close to 1 amp on each rail.
Much thanks Tad
I am already using the bulb tester. I have not turned the bias down as you suggest. I will most certainly start with the bias adjusted much lower. I had forgotten that the variable resistors do not always arrive at a predetermined value. When I went back and checked I was close to 1 amp on each rail.
Much thanks Tad
Like this:
This is portuguese, and translates like this:
Rede elétrica = MAINS
Lâmpada = Light Bulb
An externally hosted image should be here but it was not working when we last tested it.
This is portuguese, and translates like this:
Rede elétrica = MAINS
Lâmpada = Light Bulb
Tad,
I posted somewhere (In one of the Leach amp threads) that R13 + R65 should be left out - and if used should be larger (kOhms) than the 10 specified in the BOM... I can not find my original post, but andrew took part in the discussion also - Andrew can you help?
For now simply remove the resistors from your board, and lets take it from there...
Found: http://www.diyaudio.com/forums/grou...sen-leach-clone-group-buy-57.html#post2230109
\\\Jens
Last edited:
Jens,
I originally left those two resistors out of the build, as you mentioned. When I was starting the testing I was not getting any power to the output. So I soldered in the 10R resistors just to see if that was causing me not to get any output. I later found it was something else. Anyway, those little 1/2 watt resistors would not hold up to the 1 amp bias. They are now not in the circuit, or anywhere else, and I am going to proceed from here with substantially less bias.
For what it is worth 1 amp, per rail, translates to 168 milliamps per device which is not at all excessive.
I hope to have some information on this Leach 12 clone later this week. If all goes well we can proceed with the latest group buy. I would like a new set of pcb's myself if I am successful here.
To be continued --- Tad
I originally left those two resistors out of the build, as you mentioned. When I was starting the testing I was not getting any power to the output. So I soldered in the 10R resistors just to see if that was causing me not to get any output. I later found it was something else. Anyway, those little 1/2 watt resistors would not hold up to the 1 amp bias. They are now not in the circuit, or anywhere else, and I am going to proceed from here with substantially less bias.
For what it is worth 1 amp, per rail, translates to 168 milliamps per device which is not at all excessive.
I hope to have some information on this Leach 12 clone later this week. If all goes well we can proceed with the latest group buy. I would like a new set of pcb's myself if I am successful here.
To be continued --- Tad
yes, but I need to model the PSU and output stage to arrive at a de-rated SOAR and from there determine a set of protection resistors for the VI detector.
Andrew,
If you use Dr. Leach's method of bias adjustment by reading the overall rail draw at the fuse you divide that by 6 output devices per rail and you have approximately 168 milliamps. Though this is not the preferred method here at DIY audio it works for me. I know exactly what the amp load is at each fuse location.
I do not know where you got the 10R resistor from unless you are referring to the two limiter circuit resistors which I vaporized. I have yet to check what the current load is across R13,R65. I will do that in the morning. Perhaps I was just not using a stout enough resistor. The BOM states 10R.
More in the morning. Tad
If you use Dr. Leach's method of bias adjustment by reading the overall rail draw at the fuse you divide that by 6 output devices per rail and you have approximately 168 milliamps. Though this is not the preferred method here at DIY audio it works for me. I know exactly what the amp load is at each fuse location.
I do not know where you got the 10R resistor from unless you are referring to the two limiter circuit resistors which I vaporized. I have yet to check what the current load is across R13,R65. I will do that in the morning. Perhaps I was just not using a stout enough resistor. The BOM states 10R.
More in the morning. Tad
do not use a higher wattage resistor.
That will let a high current pass for longer and risk more damage downstream.
If you feel you must use the 10r in lieu of the fuse, it would be better to use 125mW or 250mW not 500mW and certainly not 5W.
But 1A through a 10r fusible link is nonsense. Set the output bias to zero. Check Vre.
If this method is not acceptable to use then use 0r1 in the fuse locations. Now 1A dissipates 100mW. you can just get away with 125mW and it only drops the supply rail voltage by 100mV to allow fairly accurate setting of the output bias. But please recognise that fuse current method is a very poor way to assess or set bias current.
Then you can check all the other operating voltages to ensure the circuit is working as designed.
Then put back the F rated fuses and power up through a bulb tester.
Check operation again.
Start advancing the bias adjustment. The bulb will start to glow. Now bypass the bulb tester and set the bias current using the Vre method.
Check that all Vre are close to each other, or at least within what you consider to be adequate tolerance.
Last edited:
I think I finally wrapped my head around how the protection circuit works, at least omitting R13 and R65. Reading other threads on protection made me curious, and I developed the attached spreadsheet to calculate the resistor values.
To date, I have just been using Leach's V4.5 protection circuit resistor values in my 3 pair amps. That gets me a protection line that is just inside the MJL3281 250 msec SOA curve. In my application, my sinks don't rise more than 10C above ambient, so derating for that rise still keeps my protection curve inside the 50 msec curve. It looks like for the three pair version, the original Leach values are a decent choice. For 5 or 6 pairs or higher SOA devices you can get more aggressive.
Somebody please verify that I got this right before too many people use it.
How to use:
Calculate load lines - using your choice of method and at least some reactive component with your worst case expected load. AndrewT posted an article yesterday in another thread that made it gel for me. I made the Load line spreadsheet attached from the article's formulae.
Plot derated SOAR curve. Whether you use 10 msec, 100 msec or DC curves depends on your application and how aggressive you want to be. PA amp or never blow up under any circumstances go DC. There is some argument over 10 and 100 msec for home use. I am comfortable with 10, you may want to be safer.
Use the spreadsheet to place your protection curve outside the load line but inside your chosen derated SOA curve. You'll input rail voltage, emitter resistor value, number of output pairs and protection circuit resistor values.
Edit: Here's a link to AndrewT's post with the articles http://www.diyaudio.com/forums/soli...sistor-safe-operating-area-4.html#post2297122
To date, I have just been using Leach's V4.5 protection circuit resistor values in my 3 pair amps. That gets me a protection line that is just inside the MJL3281 250 msec SOA curve. In my application, my sinks don't rise more than 10C above ambient, so derating for that rise still keeps my protection curve inside the 50 msec curve. It looks like for the three pair version, the original Leach values are a decent choice. For 5 or 6 pairs or higher SOA devices you can get more aggressive.
Somebody please verify that I got this right before too many people use it.
How to use:
Calculate load lines - using your choice of method and at least some reactive component with your worst case expected load. AndrewT posted an article yesterday in another thread that made it gel for me. I made the Load line spreadsheet attached from the article's formulae.
Plot derated SOAR curve. Whether you use 10 msec, 100 msec or DC curves depends on your application and how aggressive you want to be. PA amp or never blow up under any circumstances go DC. There is some argument over 10 and 100 msec for home use. I am comfortable with 10, you may want to be safer.
Use the spreadsheet to place your protection curve outside the load line but inside your chosen derated SOA curve. You'll input rail voltage, emitter resistor value, number of output pairs and protection circuit resistor values.
Edit: Here's a link to AndrewT's post with the articles http://www.diyaudio.com/forums/soli...sistor-safe-operating-area-4.html#post2297122
Last edited:
Krisfr,
Somewhere in the first group buy of this amp layout Jens remarked that these boards would probably work as a homemake pcb project. I beleive you will have to have the boards machine routed. Of course if you have one of those you are way ahead of the game. Email me and I will see if I can find the full layout for the 12 transitor amp somewhere on this computer.
Bob, How would you go about testing a set of amps layed out on the bench with the different resistor values. My boards are mounted to the heatsinks with a power supply and such which makes them fairly easy to work on.
Can this be done without a scope? With 6 pairs of outputs I should conservatively be able to drive 24 amps. This is way more than a 2-1kw trannies will drive.
Tad
Somewhere in the first group buy of this amp layout Jens remarked that these boards would probably work as a homemake pcb project. I beleive you will have to have the boards machine routed. Of course if you have one of those you are way ahead of the game. Email me and I will see if I can find the full layout for the 12 transitor amp somewhere on this computer.
Bob, How would you go about testing a set of amps layed out on the bench with the different resistor values. My boards are mounted to the heatsinks with a power supply and such which makes them fairly easy to work on.
Can this be done without a scope? With 6 pairs of outputs I should conservatively be able to drive 24 amps. This is way more than a 2-1kw trannies will drive.
Tad
6pairs can easily pass 18Aac (~25Apk) continuously as long as the heatsinks stay at or below their design temperature.
6pairs can easily pass transient currents (audio & music) of >50Apk and possibly 90Apk on very short transients.
These current peaks come from the smoothing capacitors not from the transformer, nor from the rectifiers.
Of more concern are the driver transistors.
Can they pass sufficient current to the 6output pairs to meet that 90Apk demand?
Don't try this with higher voltage PSU & 2sa1943/c5200
In a belated response to Bob Ellis’s plea for a sanity check on his spreadsheets (and since to my knowledge no-one else has taken it on), here’s my two-pen’orth.
Load Line:
(As devised by David Eather and “Excelled” by Ben Janssen)
Cell B9 should read SQRT(2*B7/B5)..
Leach protection file
The lack of the volt drop across the emitter resistor in Bob’s spreadsheet renders the locus somewhat inaccurate. I’ve corrected that in my spreadsheet. Also, I’m not sure that a negative Vce can occur at the positive side of the output stage (or is this merely to allow higher Vcc’s to be displayed?). And the Vce plot should extend all the way to the 2Vcc limit*. I’ve also cleared up some (to me) confusion over designations. Note that only three points are needed to establish the complete protection boundary with such a simple scheme.
I’ve shown Protect_Locus_2 to emphasise the wisdom of including R13 in the protection scheme (the macro for this is not shown here). This enables the horizontal section in Locus_1 to slope below the 250ms SOAR, given suitable resistor values, something that cannot be done with the original Leach scheme. Now, nowhere does the protection locus violate this curve. The load lines may, of course, but that can be avoided by increasing the number of output pairs if the same power is to be achieved.
These schemes have serious limitations at this high(ish) supply voltage since using D1 referenced to 0V makes the break point Vce invariant at around 59V (for a 60V supply). Things get much, much worse when 70-80V rails are mooted. Here a variable break point or better still a triple locus is favoured, but that will require additional component locations on the pcb.
Nothing here is new, but a spreadsheet that instantly shows the effect of varying the supply voltage and resistor values superimposed on the SOAR and load line takes a lot of the headache out of choosing a protection locus.
If you see any errors let me know!
Brian.
*Ref. Michael Kiwanuka: Transparent V-I protection in Audio power Amplifiers (EW Sep-Oct 2002)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.