Another Discrete Amp - GainBloak

[tk@halmi]

>I found the top half of the wave was being clipped & got rid of the clipping by upping the 200R resistor R111 to 300R.

You do reduce the avaialble gain a bit by increasing the emitter feedback resistor. The value also depends on your design goals. Do you want more OLG or more symmetrical clipping?

After prototyping the push-pull version I believe that the single ended output sucks. It is more trouble than worth doing. Perhaps it is good as a headphone amp, but it is not so useful in this application. There are plans to put it on a 990 footprint board and the heatsinks required are very impractical on a little board like that. The push-pull version is still Class A if you bias the transistors sufficiently. We have 6ma running in the VAS. Running that throught two diodes you get pretty good Class A bias on the output. I measured 30mA standing current one time.

Try building it on a breadboard and see how well your simulations match reality. Come on, it is only five transistors and a few resistors!

Cheers,
Tamas

 

[tk@halmi]

Here is the last installment of the GainBloke. This is the version with a push-pull output, still Class A at light loads.

http://mysite.verizon.net/res75okq/sitebuildercontent/sitebuilderfiles/gainbloke4.pdf

It is still just five transistors and seven resistors. It handles low impedance loads better than the single ended version although with light loads they have similar performance.

Cheers,
Tamas

 

[Bauman]

I present you Tamas GainBloak.

Did some quick tests in the API-ish micpres, sounds good! Need more time to have more comments.

Thanks Tamas for this cheap design!

 

[rafafredd]

Nice board Fabio... Would you test it with a low ratio transformer, like 1:1 and 1:2?

 

[Bauman]

Rafael,

tested with 1:2 jensen JT-16-B and 1:5 JT13k7A, this last one sounded better to me, in the quick test I did.

:guinness:
Fabio

 

[tk@halmi]

The open loop gain is relatively low in these amplifiers so they will struggle to produce clean gain at 40dB or more. When you use a higher turns input transformer you relieve the opamp so it does not have to work so hard.
However, they are cheap and simple so doubling them up for a dual gain stage is not going to drain your wallet. Let's say you set each stage to 30dB max, add a 1:2 input transfromer and get 66dB gain without strain.

 

[tk@halmi]

Feels like I am sitting on needles waiting for them!!! :grin:

 

[Bauman]

Hats off to Tamas.
This opamps sound sooooooooooooooooooooooo good man! :thumb:

Have tried with 1:2, 1:5, 1:7 and 1:10 input transformers... all nice. It sounds better with cap
on output than with a servo circuit...

:guinness:
Fabio

 

[Flatpicker]

[quote author="Bauman"]Hats off to Tamas.
This opamps sound sooooooooooooooooooooooo good man! :thumb:[/quote]C L! :green: :thumb:

 

[Kev]

well done Tamas
and that includes Fabio too
... it's good to know a few guys do still have the soldering iron wramed up and have there head down working away

:thumb: :thumb:

oh and hi to Pete C.

 

[tk@halmi]

Kev,

Thank You for the kind words. It could not have been done without Fabio.

Actually, the design was inspired by his comments about parts availability in South America and the mess of sorting/matching components for most discrete opamps. I was hoping to create something really easy to DIY while giving decent performance.

Well, things certainly work out better when I stick to the soldering iron instead of shooting my mouth off here and there. :wink:

Next project is a transformerless mic pre that sounds like it had a transformer. :shock:

Cheers,
Tamas

 

[Consul]

Wow, this entire time this was going on and I completely missed it?! :shock:

Fantastic job, Tamas and Fabio. I might have to give it a go myself. :thumb:

 

[Kev]

Yep,
there is often really cool stuff on the go just under the surface.

I look forward to Tamas' transformerless mic pre that sounds like it had a transformer
cos
I have a live sound job in the near future.
Too build a simple rack that will serve multiple purposes.
A simple IEM rehearsal system.
This system then feed a main PA system via a post mic-pre split (+4/-60)
Also has a simple FOH mixer for the very small gig
and also provides a multitrack record out (+4)

hopefully it will be set and forget so a simple limiter/soft clip could be in there too.

 

[Bauman]

Next project is a transformerless mic pre that sounds like it had a transformer. Shocked

:thumb:


And for the opamp, I´d say try yourself don´t listen to me. :grin:
So cheap and so good sounding you will cry.

:guinness: :sam:

 

[Consul]

Do you have a PCB layout anywhere, Fabio?

I've taken a look at the schematic, and I think I see how it works. It's kinda tricky filtering out what caps, resistors, and other bits (like the feedback loop) are there just to keep the simulation happy, and I was hoping that having a layout would help me understand it.

Thank you all for the help.

 

[tk@halmi]

Hello,

The opamp should be used in non-inverting mode. The feedback loop can be 10k and up. It can take very low feedback networks like 1k, but it makes the bandwidth very high (we are talking about AM radio band) and may cause stability issues. Since it is a current feedback topology the feedback resistance sets the bandwidth directly. However, you get slew rate and bandwidth almost independent of the gain setting. With voltage feedback those parameters change with gain setting. Use a large cap after the resistor coming off the inverting input to ground, like a 4700uF Panasonic FC, to block DC gain.
Another nice feature is that the opamp is unity gain stable without any compensation capacitors, and that is nearly impossible to do with voltage feedback.
Hang a couple of 47uF caps near the output transistors off the power rails to keep them juiced up.
The output should be capacitor coupled due to the inherent DC offset.

I am dorking with Pad2Pad to make PCBs as well. It may be a couple
of weeks before I complete it, but the per board price will be just a few bucks.

Cheers,
Tamas

 

[bcarso]

What might be fun as an enhancement: concoct an offset nulling circuit, so that at low gains the circuits could eliminate the big C's.

 

[JustinS]

Hey Tamas,

Any closer to getting some pcb's?? :grin:

Justin.

 

[tk@halmi]

Yes. Master Bauman has been busy. Give it 4 to 5 weeks. I will post in the BlackMarket section.

 

[JustinS]

Cool!

By the way I love your tagline... where is it from - Blackadder, Monty Python... current US policy??

Reminds me of a Rowan Atkinson skit "A good fatal beating never hurt anyone" :grin:

Justin.