Marshall TDA7293 substitute amp

tired of dealing with these buffalo chips,

marshall valvestate and other chip amps use them,

why not replace with a couple of power mosfets or power transistors ?  need about 80 watts,

here is the original marshall circuit, the signal has been amplified and voiced, just needs a pwr amp,

we have 4 v-rms available as input signal to whatever device we use,

only drawback we can see is short circuit protection for the speak, 
thanks for any help!

> replace with a couple of power mosfets or power transistors

Need a lot more than "a couple".

It is very hard to beat the TDA chips.

Unless you chisel the chassis clean and build a 5F6A in the hole.

what if we build a transformer to split the signal?

got some 3055's on the junk drawer,

maybe go single ended and use a big cap for speak protection,

or be smart and cobble in an LM3886 in place of the popcorn TDA chip?

we are gonna try this and see where the smoke comes out,

hopefully the marshall opamp will drive the transformer, only need about 1 ma which gets stepped down to 6 ma, enuff to drive a pair of 3055's,

the tramnsformer isolation will hopefully remove the inherant flaw in the valestate design,

marshall meets acoustic,  if it works, our chip amp nightmares are over from here to enternity.

CJ said:

... which gets stepped down to 6 ma, enuff to drive a pair of 3055's,...

Click to expand...

6 ma X hfe of 20-70 = way less than 1/2A signal at the collectors, no?

Gene

yeah i do not know whats up with the 3055 spec sheet, the transistors we picked have an hfe of 225,

did a gain plot up to 600 ma (max I for HP pwr supply) 

that acoustic used 3055's and we  have just wound the xfmr coil like the acoustic driver iron, so we will see what happens,


using 87 EI  with a 1" stack so we will have about 30% more pri inductance to make it easier for the opamp to excite the core at low freqs,

original gapped xfmr had 4.4 H at 10 H.

You could try re-creating the Marshall 2200 power amp section ,I fixed up one of these years ago for a buddy and they sound awesome,one of the first 100 watt transistor guitar amps ever made .
I found a modern equivalent output transistor and its worked flawlessly ever since
Heres a link to the schematic,
http://drtube.com/schematics/marshall/2200-pwr.gif
I dont think a Hfe of 225 for a 2n3055 sounds right by the way.

thanks for the circuit!

your right, the gain does come down as Ic passes about .75 amps,  hfe drops to 190 and plummets from there to max current,

so we might need 10 ma per base to drive it good,  so 20/6(xfmr step down ratio) = 3.3 ma from the opamp  plus excitation current for the xfmr,

which will be at 20 hz- 2 pi f l = 6.28 * 20  * 6 Henries (est)  = 125.6 * 6 = 750 ohms, 

drive voltage from the opamp is max at 4 v-rms so 4/750 = 5 ma  so we need 5 + 3.3 = 10 ma on a bad day, will check specs for opamp to see if we need a transistor on the tail end,

> Hfe of 225 for a 2n3055

Original-recipe "baked potato" '3055 had a base-width bigger than an SMD and correspondingly low hFE.

New-'3055 on planar process may have higher low-current hFE. But the old-old datasheet for '3055 shows hFE falling quite badly above a few Amps, and I think many old designs relied on this to protect (ha) against destruction. In planar, a falling hFE is also a lower-cost product, so they ape the old hFE curve when selling as a '3055.

Speaking of protection: that Marshall plan has over-current but not SOA protection. The other transfo-drive design has no protection at all. In stage amps, this means the first shorted speaker cable is an amplifier death. Much more protection was soon seen to be "essential" for products which would come back for repair at your expense.

> cobble in an LM3886

The '3886 is not good for the expected 80 Watts. If it is 80W in 4 Ohms, parallel '3886 will nominally do the job, but you need heavy heatsinking or they will self-protect in the middle of your brilliant solo. I don't see how this is better than '7293.

those chips are fine if you do not overwork them, but there appears to be a problem with this discontinued amp whereby it is unstable and so when you replace the chip, it works for a while and then dies in a puff of smoke, sure enuff, when we put a new chip in there from digikey it started bubbling like a tar pit, might be a plastic part,  so we are trying to find something more bullet proof,

apparently we ar not alone with this problem>

"I had an AVT50 with the same problem, was totally unstable.
The amp was just out of warranty and had been sent back for repair twice for the same problems.
After trying everything to repair the amp I eventually replaced the Marshall power amp PCB assembly with an off the shelf kit with similar specs and a decent finned heatsink.
All the nasty distortion and HF oscillation dissapeared.
I am convinced that this particular model has an inherent design fault, the output chip has less external support components than the manufacturer's recommend to make the original device used stable.
Don't be surprised if your friend's amp fails again."

this amp works fine, no feedback or ultrasonnics until you hook the pwr wires to the chip, then all hell breaks loose at hi gain settings, we are thinking there might be a ground issue with the pwr daughter board, we are hoping that the innerstage xfmr provides the isolation we need to fix the feedback and make the amp last longer due to better heat sinking from the twin TO3 package,

Ive worked on alot of the Marshalls over the years and I have to say the Avt series arent the greatest, the first generation valvestate gear wasnt too bad aside from the odd component breaking off the board . Of the modern stuff the MG-100 is fairly rock solid ,it uses two 7293 modules to make 100 watts into 4 ohms . Like most transistorised guitar amps if you end up clipping the output stage bad things tend to happen. A guitarist I know had an Avt275 which he finally got rid of after struggling for years to get a sound out of it ,something about the stereo poweramp set up in them and a fairly marginally spec'd powersupply just didnt do the business at higher volumes . I see now marshall have ventured into domestic audio products,I guess the marketing men have taken over the operation now Jim has passed away.

most of the bone pile here is all chip amps, dan electro, fender, behringer, now marshal, if we can come up with a low parts count way of making them bullet proof, then we will have 127 cubic feet of room for tube amps, 

using an old transistor for a drilling template,

you can not drill the blue glass epoxy that seals the base amnd emitter leads, but a big enuff hammer and center punch makes it shatter like a dry twig in a stiff desert wind,

looky here: how to wind your own driver transformer!

http://www.geofex.com/Article_Folders/Thomas%20Vox/Vox%20Driver%20Transformer.pdf

I know it's not DIY in the strictest sense, but I built one of these from the kit:
http://www.ampslab.com/hx100.htm
100 W into a 2x12" 4 ohm cab on guitar all day long

and one of these from the kit:
http://www.ampslab.com/bi120.htm
100+ W bass into a 12" horn load 4 ohm, just gets warm, hell it even includes the nice heatsink

first one is like $100, takes an hour to put together, what's your time worth?

Joe

thanks for the links! mosfets seem to sound better for guitar/bass.

those are handy for a bench pwr supply, plug your generator in there and you can check output transformer saturation at 10 Hz,,

money? more of a hobby, so i have more time than money,

got the driver coil done, from that DIY driver transformer article it is good that we used thick wire for the secondary as base drive current needs to go up when cranking out max amps,

going to slap a transistor on pin 8, so three transistors, six resistors and a transformer will hopefully be our parts count, oh, and a chunk of turret board from King Kong,

notice the hi-tec mandrel from a 57 plymouth,

Yeah its is a cool article on transformer winding ,Ive been meaning to try experimenting for a good while now ,there are a few nice looking coil winders on ebay out of China ,not hugely expensive either.Tube interstage or neve style output would be probably similar to what the guy did in the article.

coming along on the chip amp replacement program,

add 1 peg board plus Acoustic 150  innerstage TX>

stuff borad with pwr amp components and measure, we have a little less voltage than the Acoustic, about 62 instead of 75, no big,

TL072 a no go for driving gapped TX, not enuff pri Z,

add rest of Acoustic driver circuit, (bot left) scale AC voltages later,

TIP 30 heat sinked to chassis, sub for RCA heatsinked 40410, MPA A06 drives TIP,

driver gnd has to be changed, output transistors run off of  plus and minus rails, too much hum to run driver SE driver, must change gnd to chassis (CT of pwr supply)

less voltage but we still can rail out 3055's without distortion from driver

3055's heat sinked to old chip amp chunk,

no mucho heat when cranked, 

driver circuit cool as a cucumber with reduced single rail supply, 

sings like a bird with no more feedback, UL osc and melted chips,  ;D ;D ;D    xfmr did the trick, adding iron is always good,  ;D

sounds like a Neve guitar amp,