Any Maplin MosFet Amp Guru's on here?

[davidsrsb]

Mine have K133 and J48 and have been running at +/-63V since I built them in the 80s

120V max devices on 126V total rail - not recommended
The trouble is that a blown output transistor has a habit of blowing the speaker woofer very quickly

 

[traderbam]

It’s ok... his amps have a closed loop gain of 1 so they don’t get very loud.

 

[traderbam]

Looks as though a re-layout is in order, In Line Fuse holders are a good idea, I'll have a look and see if I can find some that would take the gauge of wire I'm using, possibly something in the Automotive line.

Your wire is very thick.

 

[DiyGazza]

Your wire is very thick.

Not overly thick , It's 4mm 32A Stranded Mains Cable, you can see in the pic, post #19.

 

[davidsrsb]

The major defect in the Maplin design is the lack of gate clamp. It really needs a couple of zeners to prevent the mosfet gates breaking down when the output is shorted. Lateral mosftets are too expensive to zap these days

 

[DiyGazza]

Maybe I should just bin these and build something better, Any suggestions ?

 

[davidsrsb]

The core design works well, but needs a few tweaks to be more reliable.
For example, the laterals are specified as "cut off" 100mA 0.15V<Vgs<1.5V
The voltage across RV1 should only be able to reach 3V at maximum, so 500R is right, not 1k

 

[davidsrsb]

The V2 schematic

 

[DiyGazza]

The core design works well, but needs a few tweaks to be more reliable.
For example, the laterals are specified as "cut off" 100mA 0.15V<Vgs<1.5V
The voltage across RV1 should only be able to reach 3V at maximum, so 500R is right, not 1k

I agree, these really are nice sounding little amplifiers, and if a few minor tweaks are necessary to make them a little better, then so be it, I fully intend to go with the suggestions given by "traderbam", but I really don't want to go to the expense of replacing the toroid.

It would be nice to see some form of protection for the outputs, but one thing that I could never make sense of is the fact that you run a nice big cable from your amps to the outputs, and nice thick cables to your speakers, but then stick a relay in the path with its contact resistance, and if you are unlucky, which has happened to me on several occasions with different amps, the contacts oxidize and start to crackle.

 

[DiyGazza]

@ traderbam

Would these be suitable replacements for C8 and 10?

Not Suitable for across the line application?

R75MR41004040J | KEMET 1μF Polypropylene Capacitor PP 220 V ac, 400 V dc +-5% Tolerance Through Hole R75 Series | KEMET

 

[scottjoplin]

The major defect in the Maplin design is the lack of gate clamp. It really needs a couple of zeners to prevent the mosfet gates breaking down when the output is shorted. Lateral mosftets are too expensive to zap these days

The Hitachi Mosfets used have internal protection diodes

 

[nigelwright7557]

The major defect in the Maplin design is the lack of gate clamp. It really needs a couple of zeners to prevent the mosfet gates breaking down when the output is shorted. Lateral mosftets are too expensive to zap these days

The zeners would also limit current through output mosfets.

I have a lateral mosfet application note that recommends that.

 

[davidsrsb]

The Hitachi Mosfets used have internal protection diodes

The original Hitachi parts did, but at ~14V, too high a voltage to limit the current to a safe level
The more modern equivalents from Semelab etc do not have the zeners and you won't find genuine NOS Hitachi parts easily

 

[nigelwright7557]

The original Hitachi parts did, but at ~14V, too high a voltage to limit the current to a safe level
The more modern equivalents from Semelab etc do not have the zeners and you won't find genuine NOS Hitachi parts easily

Some great lateral mosfet application notes here.
They also mention zeners to limit output current.

http://www.ckp-railways.talktalk.net/alf.pdf

 

[davidsrsb]

Thanks for the link Nigel. The laterals seem to be going out of production, so proper device proection is becoming more urgent. (Could get rid of those fuses too)

Another reason to use the right value of potentiometer to set quiescent is that if the wiper goes open circuit, the current goes to maximum. With 3V max and 2V typical, the current might be about 1A, survivable for a few seconds.
6V max means about 3A, not enough to blow the fuses, but 150W will melt the device in seconds

 

[nigelwright7557]

Thanks for the link Nigel. The laterals seem to be going out of production, so proper device proection is becoming more urgent. (Could get rid of those fuses too)

Another reason to use the right value of potentiometer to set quiescent is that if the wiper goes open circuit, the current goes to maximum. With 3V max and 2V typical, the current might be about 1A, survivable for a few seconds.
6V max means about 3A, not enough to blow the fuses, but 150W will melt the device in seconds

I always put a zener across the potentiometer to catch a faulty pot.
The voltage is just a little higher than the normal voltage across the pot.

 

[Xoc1]

There are many variations on the basic Hitachi design
I at one time it seems that half the British PA amplifier industry was using lateral mosfets.
One of the first would have been the H H V800
Schematic
H|H Audio V800 Amplifier Information
This is very basic but at least has the zener diode protection for the gates.

Other companies would include
BK electronics - basic circuit but with CCS for the input pair.

I have a Macgregor 500W mosfet amp that is the same basic circuit but polarity reversed witn PNP and NPN swapped!

Rauch DVT series amplifiers are interesting
The first DVT amplifiers used the most basic circuit - very much like the Maplin.
Later versions added a constant current souce for the input pair.
Then a proper driver stage was added for driving the gates.
I modified a Rauch DVT250M - 1100W@2R monoblock - This had 16 mosfets running on the basic app note circuit. It was a lot more stable with a proper driver for the gates added.
Zener limiting for the gates was also eventually added. Before the zeners were added a short to the outputs would cause the DVT amplifiers to blow up in a spectactular fashion

The most complicated version I know of would be the C-Audio SR and SRX
Therse have CCS added for the Input stage CCS for the VAS second stage and a driver stage that is bootstrapped.

Schematic
C Audio sr707 input pf capacitors issue

So the question is how far do you want to go with mods!
Regardless of the variations with the circuits, instability caused by oscillation of the mosfets is the bane of these amplifiers.
Cures have included;-
increased resistance values for the gate resistors
Miller caps from Gate to Drain
Ferrites on the gate resistor lead
and even individual zobel networks from Source to Drain on all the FETs

 

[DiyGazza]

So How easy is it to add a zener Gate Clamp to help protect against a short on the output, what voltage zener, it's current rating etc, the same goes for a zener across the Bias pot, what would we be looking at here.

What you need to understand is, I'm no designer, Yet I can repair a an aircraft flight computer, Head up display, etc with no problems, but the circuits have already been designed, I Know what signals are going in and I know what I should be seeing at the other end, so its not too difficult to work through the circuit to work out where things are going wrong, and I have been doing this for the last 30 years and up until recently where illness forced me to leave work.

Before that, I worked for a company that built Amstrad and Fidelity audio units (yeah, I know) , I also had a period where I worked for myself repairing PA Equipment for local and not so local bands, Hundreds of blown 2N3055's, lots of Marshall amps where the Tubes glowed so brightly they lit up the room, due to the X Class Cap which supplied the -42 Bias voltage going O/C, and also TV and Video - (Shows how long ago that was) !!

But, give me a formula, and it might as well be written in Chinese, the same goes for Code, Macros etc, it's one of those things I can stare at for hours, and try to understand, but it just doesn't register.

So as people start to suggest adding zeners to help protect high gate voltages, that's fine, but do they simply connect between ground and the gate, or would there be additional circuitry involved, and as mentioned what voltages would we be looking at?

I hope that some you can understand my dilemma.

 

[nigelwright7557]

It depends on the mosfet resistance.
The mosfet spec I uploaded earlier is 1ohm SD reistance so its 1 volt per amp. So to limit current to 10 amps means a 10 volt zener.

 

[davidsrsb]

The modern 8A and 16A parts both hit maximum current at about 7V, hence the 6V8 zener plus a diode in the HH design

Gate resistors should be in the 100 to 300 Ohm range. The critical layout is that the resistor should be no more than a few cm away from the Gate. Too many designers used the resistor as a convenient track jumper and left a long track. This makes a nice VHF oscillator.