Welcome to Solid State Guitar Amp Forum | DIY Guitar Amplifiers. Please login or sign up.

March 28, 2024, 05:59:55 PM

Login with username, password and session length

Recent Posts

 

Randall RG100 hybrid

Started by ENGR BUGO aka DimebuGG, October 12, 2011, 12:37:45 PM

Previous topic - Next topic

allan_belt

Quote from: J M Fahey on May 01, 2012, 12:03:02 AM
No problem.
I have posted it earlier myself, but don't remember the link, so repost.

EDIT: I saw the project you found.
Don't like it very much: it uses the obsolete single supply configuration, has no short circuit protection, the feedbark network is unnecessarily complex (and probably wrong) and it's not easy to modify it into a mixed feedback system.
Study this:
http://www.diyaudiotr.com/islecler/tammetin.php?metinid=6&durum=e#sonyorum
It is *very* crude, simplified to the maximum, but it works and can be built.
The PCB is easy to homebuild and works too.
And is very easy to modify to mixed feedback.
It's in Turkish but google can translate the page easily.
http://translate.google.com/translate?sl=tr&tl=es&js=n&prev=_t&hl=en&ie=UTF-8&layout=2&eotf=1&u=http%3A%2F%2Fwww.diyaudiotr.com%2Fislecler%2Ftammetin.php%3Fmetinid%3D6%26durum%3De%23sonyorum
Some small corrections:
1) forget the 150W they claim and even 100W, it can provide honest and loud 70W into an 8 ohm speaker.
2) the supply voltage suggested is not too safe.
I suggest (by experience) not passing +/- 40V rails, which demand a transformer of no more than 30+30VAC (or 60VCT).
3) there is a separate .pdf with the ready to print PCB.
www.diyaudiotr.com/resim/proje/kapibara/pcb.pdf
I recommend the iron on laser thermal transfer process, but it requires some experience.
Meaning, make 3 or 4 PCBs before trying this one.
I suggest not building this amplifier now, but it's good to know what you are getting into.
Start by making an LM386 amp.
Build and enjoy it. It is fun.
Later, build , say, some 15 to 30W with a TDA2003, TDA2030 or even TDA2050.
Build and enjoy.
Later, an LM3886 project. (50W)
Loud enough to play anywhere.
Go step by step, enjoy and learn along the way.

I see you live in Peru. Check what amplifier kits you can buy in an Electronics Shop.
Hello again , i wasn't too happy with the ic based power amp sound so i decided to build the power amp above based on Darlington transistors .. and the results were better than with the TDA7293 power amp.
Louder and punchier ..
Here the original schematics:


now my questions are:
if i want to mod it to mixed feedback is the following correct?
and what would be the ideal values ?


i really appreciate your help. Cheers from Peru.


J M Fahey

Basically right, but please repost the schematic labelling every component , such as R1/R2 ....C1/C2.... Q1... etc.
Also label input/output and +/- on the speaker.
I *hate* explaining "the resistor to the left of the transistor which is upside down between ..... "
Also the added feedback resistor goes to the lower end of the speaker, not straight to ground.
Good luck.

allan_belt

#32
Quote from: J M Fahey on September 15, 2012, 12:36:59 AM
Basically right, but please repost the schematic labelling every component , such as R1/R2 ....C1/C2.... Q1... etc.
Also label input/output and +/- on the speaker.
I *hate* explaining "the resistor to the left of the transistor which is upside down between ..... "
Also the added feedback resistor goes to the lower end of the speaker, not straight to ground.
Good luck.


Whoops i made a mistake in the drawing  like you said. Now here is the corrected schematic:


thank you ... :dbtu:

allan_belt

Any help please??? i need the values of the new components..  ;).

I think this will be my last try in power amp section then i`ll post mp3 demos  :tu:

Loudthud

#34
Break the connection of C1 to ground and connect it to the top of R12. If you want to experiment, connect a 25 ohm wirewound pot (1 or 2W) across R12 and connect C1 to the wiper. That way you can adjust between voltage mode and mixed mode feedback. There will be some gain change as you adjust the pot. R12 can be 0.22 5W (or higher). JM basically gave you the answer in post #21.

allan_belt

#35
Quote from: Loudthud on September 18, 2012, 01:04:23 PM
Break the connection of C1 to ground and connect it to the top of R12. If you want to experiment, connect a 25 ohm wirewound pot (1 or 2W) across R12 and connect C1 to the wiper. That way you can adjust between voltage mode and mixed mode feedback. There will be some gain change as you adjust the pot. R12 can be 0.22 5W (or higher). JM basically gave you the answer in post #21.
Cool thanks Loudthud , so C1 and R3 should instead of C5 and R9? yes i think i need to read more about Mixed mode feedback.. concerning to the pot instead of R12 i think i'll keep R12 at 0.22 Ohms .. ;) (mixed mode feedback) because i`m trying to get close to the Randall RG100es Power amp:

by the way i see in Randall's schematics R44 and C33 why are they there?  ???
Would be a good idea to put the same values of the Randall in my Darlington power amp (i mean only R44, C33,C36, R52 and R63)? thanks for all your help ..you're so nice people.. 8)

Loudthud

Randall's power amp is not what you would call "mixed mode" feedback, it's current only feedback with DC correction. C34 blocks any AC voltage feedback, only allowing DC and very low frequencys to pass. The current feedback signal from R63 gets divided down by R52 and R44. C36 and C33 block any DC feedback through that path that would otherwise kill the DC coming back through R45.

This same type arrangement could be used with a Chip or other power amp if desired, although stability might become an issue. C32 is a fairly high value limiting the high frequency open loop gain of the Randall amp.

allan_belt

Quote from: Loudthud on September 20, 2012, 11:47:01 AM
C32 is a fairly high value limiting the high frequency open loop gain of the Randall amp.

I'm confused about C32 ..  ???

Roly

Let's take a couple of steps back to the circuit in post #32.

Negative feedback, NFB, is applied to the input differential pair via R4.  If this was all that happened then the overall gain would be unity because the input and output voltages have to match in the diff pair.

Now we introduce R3 and create a voltage divider, and if we reduce negative feedback we increase gain.  The gain is now approximately R4/R3, 22000/470 = 46.81 times.

If this was also the DC gain of the amp it would tend to be thermally unstable, so we introduce C1 which effectively makes R3 open circuit for DC and the gain therefore falls to 1 for DC.

In round numbers, with +/-40V rails the peak output voltage will be about 40 volts, and if the AC gain is about x40 then the input for full output will be about a volt peak (a bit less actually), so the signal voltage at the base of Q2 will be the same.

Now in mixed-mode feedback you typically have about 10% current feedback and about 90% voltage feedback (i.e. via R4), so if the voltage feedback is a bit less than a volt peak you want about 100mV peak of current feedback at the base of Q2.

Your C5 does much the same job as C1 so I'd start off with a similar value.

Now you need to apply Ohms Law to your values for R9 and R12 so that you develop 100mV across R3/470r when you have full output.  This could get tricky without having some gain in this path because you want to generate at least ten times as much voltage so that your added R9 is large compared to R4/22k so it doesn't upset the AC voltage feedback conditions, but you also want to keep R12 as small as possible so you don't waste your output power heating it up.

Me, I'd be inclined to go with a value of R12 that gives you about 100mV for full output, then use an op-amp to bump this right up to say 10 volts (i.e. x100), then you can have a large value for your R9 to bring it right back down again without upsetting the relationship between R3 and R4.  This will also let you adjust the amount of current feedback by changing the op-amp gain and not need a power pot at R12.

The amp in post #35 doesn't seem to have AC voltage feedback at all due to the bypassing/filtering action of R46/C34.  This helps to explain why the Miller capacitance C32 is so large; the loop may well be unstable at high frequencies without overall AC NFB because the loop gain will be close to the open loop figure, which will be very large; so they apply heavy local NFB to that transistor.

R44 and C33 serve the same function as in the circuit in post #32, except it is only dealing with the feedback voltage arising from the speaker current, i.e. current feedback.

HTH.
If you say theory and practice don't agree you haven't applied enough theory.