diyAB Amp The "Honey Badger" build thread

[danielwritesbac]

Grossly approximate for class AB: VA*0.6=Watts
That is assuming 60% efficiency and most other factors disregarded.

But we might also want to look at audiometrics--"kclusa@yahoo.co" reduced bottleneck by upgrading nfb-shunt cap size for bigger deeper clearer high-resolution bass. He used a Panasonic FC 220u for C3 and an identical Panasonic FC 220u for C4, resulting in a super low loss 440u sized NFB-Shunt cap. This is twice as big and allows bigger lowest bass notes to get through the amp. You might also want to check your Arcam.

[bkevin]

Hello everyone (is there really anyone out there?)
I’ve decided to go ahead with the build – wish me luck! One problem I’ve bumped into is matching Q10,11 and 12. Best I can get for Q10 is SA1381E but for Q11 &12 (and Q13) is SC3503D. In other words they can’t be gain matched as far as I can see, or if they can it would be at a relatively lowly hfe of 100-120 (with a lot of luck thrown in finding matches!). I read that MJE340/350 could be a suitable alternative, and they seem to have a potentially better gain? Can anyone confirm or advise how to overcome this problem?
Many thanks in advance

[kclusa@yahoo.co]

Originally Posted by bkevin

Hello everyone (is there really anyone out there?)
I’ve decided to go ahead with the build – wish me luck! One problem I’ve bumped into is matching Q10,11 and 12. Best I can get for Q10 is SA1381E but for Q11 &12 (and Q13) is SC3503D. In other words they can’t be gain matched as far as I can see, or if they can it would be at a relatively lowly hfe of 100-120 (with a lot of luck thrown in finding matches!). I read that MJE340/350 could be a suitable alternative, and they seem to have a potentially better gain? Can anyone confirm or advise how to overcome this problem?
Many thanks in advance

I don't do match at all.
How do you perform match ?
Hfe changes on bias, temperature and even frequency.
If you are using the hfe function on simple DMM, That only match one point at that temperature at that current.

Lou

[bkevin]

Originally Posted by kclusa@yahoo.co

I don't do match at all.
How do you perform match ?
Hfe changes on bias, temperature and even frequency.
If you are using the hfe function on simple DMM, That only match one point at that temperature at that current.

Lou

Well that’s one way of overcoming the problem I suppose. I can’t do much better than a simple DMM test, but to my mind if they at least match at the same temp and current there’s a better chance they will match later on, and hopefully better than no matching at all. The point I was making and hence my request for help is that SA1381E is a different gain rating to SC3503D. In other words the gain match is inherently bad to begin with.

[ostripper]

Now that I have these awesome boards, all will be easy..... my experience in sourcing the right "stuff" at the lowest price will be valuable.(I will share)

My first impressions are that these boards are top quality. The resulting amp will dependent on what components are chosen for it.

The gain match of the 1381 and the 3503 is futile. the 3503's are just doing "CCS work". The 1381 is the "star player" .. It's gain (along with it's driver) , is what you hear. 10K gain + (close to) no load on the cascode is the "trick" that the "blameless" topology utilizes to achieve near zero distortion.

PS - on a current sourced amp , the CCS is a fully separate entity . You could use lamearse devices for the CCS , as long as the darlington- active (1381-ksa992) pair is there. 10K+ of gain +NFB is what makes the" blameless" rock!!
OS

[bkevin]

Originally Posted by ostripper

Now that I have these awesome boards, all will be easy..... my experience in sourcing the right "stuff" at the lowest price will be valuable.(I will share)

Excellent! That’d be a great help to amateurs such as myself in identifying the best quality components. Look forward to seeing the outcome.

My first impressions are that these boards are top quality. The resulting amp will dependent on what components are chosen for it.

Good to hear – I’m now close to pushing the “buy button”

The gain match of the 1381 and the 3503 is futile. the 3503's are just doing "CCS work". The 1381 is the "star player" .. It's gain (along with it's driver) , is what you hear. 10K gain + (close to) no load on the cascode is the "trick" that the "blameless" topology utilizes to achieve near zero distortion
"

Thanks for that. So for the 1381 you’d recommend the highest gain available? SA1381E (hfe 100-200) is best I can find - is that good enough? Or would the MJE340/350 combo (hfe 30-240) be a better bet?[/COLOR]

PS - on a current sourced amp , the CCS is a fully separate entity . You could use lamearse devices for the CCS , as long as the darlington- active (1381-ksa992) pair is there. 10K+ of gain +NFB is what makes the" blameless" rock!!
OS

[ostripper]

by Bkevin -Thanks for that. So for the 1381 you’d recommend the highest gain available? SA1381E (hfe 100-200) is best I can find - is that good enough? Or would the MJE340/350 combo (hfe 30-240) be a better bet?[/COLOR]

Either a high gain 1381 OR a high gain 992 will suffice. The ideal is for BOTH Hfe's to be as high as possible.
200(1381-F or E) x 400(ksa992-F or E)=80k. 8ma/80k = about a .1 uA(.0000001A) load on the current mirror. In real life this might be higher, (flux ,the board,or loading by the drivers).

OS

[naf]

Originally Posted by ostripper

Either a high gain 1381 OR a high gain 992 will suffice. The ideal is for BOTH Hfe's to be as high as possible.
200(1381-F or E) x 400(ksa992-F or E)=80k. 8ma/80k = about a .1 uA(.0000001A) load on the current mirror. In real life this might be higher, (flux ,the board,or loading by the drivers).

OS

Thank you for explanation

[danielwritesbac]

OStripper,
What do you think of this proposed and minor collection of updates?:
http://www.diyaudio.com/forums/attac...oneybadger.gif
There's a few different devices, much easier matching, one different cap value (for cleaner deeper bass) and hopefully non-explosive, constrained trimmer settings. Are there any errors on that schematic? Can we indeed dial the potentiometers shown to Any setting and not burst the outputs, or do those potentiometer variables need reworked again?

[ByronInLawrence]

Daniel's updated schematic says "no ON Semi small or medium signal devices -- Please use Fairchild and/or Toshiba." for 4R loads. I'd like to be able to drive 4R loads. My speakers are rated as "6 ohms nominal, 4 ohms minimum". I already purchased the ON Semiconductor NJW0302G and NJW0281G, both rated at 150W. My transformer is 800VA, 35V secondaries, so I could potentially drive them beyond their 450 watt (150 X3) rating. But then, I don't play music very loud. Is it worth getting higher rated output transistors?