Germanium Amplifier

For the three and more legged things
User avatar
Mike H
Amstrad Tower of Power
Posts: 20157
Joined: Sat Oct 04, 2008 5:38 pm
Location: The Fens
Contact:

#121

Post by Mike H »

Already sourcing it with the switcher box pre (7 metre long connecting leads), not got a valve pre as still not quite sure how to do it yet.


Anyway appendix update re the power supply ~ added a 1uF suppressor cap with mains transisent suppressor (varistor) in parallel across the mains switch, to stop the terrifying *POP* noise from the speakers when it's turned off. This is due to picking up and amplifying the back EMF pulse from the mains transformer. Also, things like clicks and pops if e.g. a light switch goes off elsewhere in the domicile, i.e. mains bourne noises
 
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
User avatar
Mike H
Amstrad Tower of Power
Posts: 20157
Joined: Sat Oct 04, 2008 5:38 pm
Location: The Fens
Contact:

#122

Post by Mike H »

Updated diagram ~ some transformer pin numbering was wrong way round:

http://livinginthepast-audioweb.co.uk/i ... 2012-4.gif

Plus, transformer connection diagram ~

Image
 
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
User avatar
Mike H
Amstrad Tower of Power
Posts: 20157
Joined: Sat Oct 04, 2008 5:38 pm
Location: The Fens
Contact:

#123

Post by Mike H »

 
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
User avatar
Mike H
Amstrad Tower of Power
Posts: 20157
Joined: Sat Oct 04, 2008 5:38 pm
Location: The Fens
Contact:

#124

Post by Mike H »

Got the new transformer installed, much better than the original ~

Image
 
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
User avatar
Mike H
Amstrad Tower of Power
Posts: 20157
Joined: Sat Oct 04, 2008 5:38 pm
Location: The Fens
Contact:

#125

Post by Mike H »

Small panic-ette today, while I was swapping speaker leads between this and breadboard amp 'tother day I was wondering why disconnecting / connecting to this amp was making alarming crackling noises and mahoosive cone movements :!:

Measure DC on speaker terminals with no loads connected there is -0.7 - 0.8V across them ~ negative ~ hmmm .. Image

Where the hell is that coming from ~ and how .. Image

Very odd.

On a hunch, switch DMM to AC, > 10V! :?: :!:

Nowt else for it, dust off the 'scope (does it still work?) and have a poke about.

Bleeding thing is oscillating, but OK while a low impedance load is connected. Very difficult to resolve what the freq. is as it's a right hash, but seems to have a fundamental 50 kHz. -ish.

It's approx. a 15V peak square wave, 50 kHz plus all the other garbage mixed in, bluddy Nora no wonder the speakers are crackling!

A load damps it so it works OK, but not really ideal to rely on that.

47R resistors across the terminals works, but, messes up the sound by noticeably interfering with the speakers performance, so that's no good.

Putting snubbers across the terminals (e.g., 10R in series with 100nF for example) just cleans up the square wave and lowers its frequency! Still coming out at 'full power' (nearly destroyed the 2W 10R in the test snubber in just 2 - 3 seconds)

I know what this is, as I've come across it before ~ if the o/p transistors have no load to speak of then the driver is also working into a higher impedance (negligible base currents, in this case), making it unstable. The transformer coupling is not ideal so it can't keep a lid on the HF acrobatics.

The solution (I hoped), is a snubber across the transformer primary, 1k & 10nF seems to work. This damps the ringing on the primary as well.

NB: also turned out this ringing started showing up as a bit of ripple on the sine wave when turned up > 10V peak.

The caps started as HV ceramic disc and then believe it or not I can hear the blasted things ("zzz - zzzz - zzz - zzz" which should instead be "sss - sss -sss")

Swapped for polystyrene, sorted.
 
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
User avatar
Mike H
Amstrad Tower of Power
Posts: 20157
Joined: Sat Oct 04, 2008 5:38 pm
Location: The Fens
Contact:

#126

Post by Mike H »

The saga continues!

Been fiddling with this most of today.

We put it on Andrew's PC based test facility at Owston which revealed some odd behaviour. Well I kind of suspected.....

OK so first prob, the HF instability. Started by disconnecting all the small value caps (get back to basics) ~ on Sat afternoon (at Owsters) I remembered an earlier version sim of this many moons ago, where a cap between mosfet drain and the NFB point at 1st transistor emitter (across what is now 100k resistor R6) could cause it to oscillate, but that was without an 8R load connected. (Instead, 470R.)

All subsequent sims were ~ yes, with 8R load connected!

Conclusion ~ MUST NOT strap a cap across the 100k resistor. (R6. This is the local NFB for the front-end.)

Playing with sig gen and scope proved that ANY caps applied to anywhere to reduce bandwidth of the front-end is a no-no, this exacerbates the problem ~ the problem being, it really needs to make the transformer work at HF, so every attempt to reduce its bandwidth just increases the ringing on the transformer, i.e. it's 'losing control'.

The resonant freq of the transformer is the peak we saw on the response on Andrew's screen, about 50 - 60 kHz.

After the peak, output runs to zero at about 100 kHz.

The resonance peak can be killed off by a rather aggressive Zobel across the primary, it ended up as 220nF & 470R.

As the o/p impedance of an amplifier rises with freq., & proved on the scope, this can make it unstable if load is disconnected, so I then experimented with Zobels across the o/p terminals until both channels continued to show a flat line AND take an input signal with the load resistors disconnected. Result. :thumbleft:

This just leaves the one NFB cap, C6.

The hum pick-up issue is most prominent on left channel ~ can't remember which one we tested ~ this is quite likely as this IT is closest to the mains transformer.

RH channel shows about 1mV peak of hum ('scope goes down to 2mV/cm, [with x2 control]), LH channel is a smidgeon over 2mV.

I had an idea Image move the IT. Undid it's nuts and moved it further away. Found a place nearer the front where the hum on both channels is the same, 1mV. That'll do, drill 2 holes there, proper job. I can live with 1mV peak of hum rather than remove & relocate the mains TX externally.

Plug it together for a listening test, OK, but it had lost its sparkle. So the IT Zobel caps were reduced to 47nF to get it back. This means it may still be prone to be unstable if loads disconnected, but so be it.

Schematic link (large)

http://livinginthepast-audioweb.co.uk/i ... 012-4d.gif
 
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
User avatar
Mike H
Amstrad Tower of Power
Posts: 20157
Joined: Sat Oct 04, 2008 5:38 pm
Location: The Fens
Contact:

#127

Post by Mike H »

PS: L4 merely represents the twisted wire pair connecting board to o/p terminals.
 
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8862
Joined: Mon May 21, 2007 9:42 pm

#128

Post by Paul Barker »

Mike H wrote:
Schematic link (large)

http://livinginthepast-audioweb.co.uk/i ... 012-4d.gif
So how does it sound say compared to a push pull valve amp? the Naime Clone or a Naim for that matter? The Gainclone? The Tripath?

Thanks?
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
User avatar
Mike H
Amstrad Tower of Power
Posts: 20157
Joined: Sat Oct 04, 2008 5:38 pm
Location: The Fens
Contact:

#129

Post by Mike H »

Couldn't really say that there was much difference, but then there were loads of different speakers and wotnot being swapped around which has the greatest effect. I only got a chance to try it with Colin's speakers (2 different)

As Nick said on Saturday we seem to have reached a plateau as far as amplifier quality goes, so the main noticeable differences were the transducers and such.

It's probably more like a valve amp in terms of open loop gain is low, so damping is low, and output Voltage is variable with impedance. Which is what I wanted, in case anyone was wondering :D

It's very similar to the breadboard 6V6 pentode amp, except the 6V6's 'don't do' bass, and slightly more mellow possibly. That amp uses Schade (resistor) feedback else it sounds too hard.

That doesn't really help does it?

Image
 
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8862
Joined: Mon May 21, 2007 9:42 pm

#130

Post by Paul Barker »

That tells me it is good.
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
Andrew
Eternally single
Posts: 4206
Joined: Thu May 24, 2007 2:18 pm

#131

Post by Andrew »

Result :!: I'm glad the session with analyser was useful.

Andrew
Analogue, the lost world that lies between 0 and 1.
User avatar
Mike H
Amstrad Tower of Power
Posts: 20157
Joined: Sat Oct 04, 2008 5:38 pm
Location: The Fens
Contact:

#132

Post by Mike H »

yes Andrew very handy, ta muchly. :thumbleft:

The odd harmonic stuff we were seeing at the top end could have been due to the o/p drops significantly once well into the ultrasonic, at which point the front-end is working harder trying to get the transformer o/p up, but which it won't.

@ Paul yes I think it is, it now does a fairly accurate square wave anyway, if with a bit of ringing on the leading edges.
 
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
User avatar
IslandPink
Amstrad Tower of Power
Posts: 10041
Joined: Tue May 29, 2007 7:01 pm
Location: Denbigh, N.Wales

#133

Post by IslandPink »

I thought it sounded very interesting Mike, from the short listen I had . I think I was in and out of the room at that point, bringing in some things, but I thought it was definitely different, maybe there was some good tone compared to silicon amps . Please let us hear it again at Owston .
"Once you find out ... the Circumstances ; then you can go out"
User avatar
Mike H
Amstrad Tower of Power
Posts: 20157
Joined: Sat Oct 04, 2008 5:38 pm
Location: The Fens
Contact:

#134

Post by Mike H »

Sure nuff.

To maybe help answer Paul's question as well, I'm pretty sure a lot of what is does is based on choice of the resistors and caps. To wit, Arcol CC & Panasonic FC elecs. The semiconductor bits seem entirely transparent.

I find it quite interesting that the OC35's can apparently be made to work as good as any more modern silicon equivalent, which suggests the old amps only sounded the way they did (whatever that is) because of the passives?
 
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
User avatar
Mike H
Amstrad Tower of Power
Posts: 20157
Joined: Sat Oct 04, 2008 5:38 pm
Location: The Fens
Contact:

#135

Post by Mike H »

Changed over to the Toco Cosmos pot. Originals were the Rapid singles, which are not too bad really.

Finally got in a proper listening sesh after tea 'til about 1 am.

After the 6V6 amp, tweeters had to be reconnected and restored to 'full power'. Means, what they had always been originally set at for months/years.

This might imply the Cosmos is 'mellow', it isn't, an apt description might be 'full-bodied'. As the reviews said bass is all there, BUT, interestingly, is still quite tight, i.e. markedly 'non-flabby'. Even the deep heavy stuff. Which was a surprise.

The treble is not 'in yer face', but nevertheless all in there. No missing bits.

This amplifier is turning into a benchmark against which anything else I build must be compared. :shock:

Played a varied bunch of stuff (in the time available), and all was ecstasy. I was on cloud nine, I was.

Chuck anything at it, it plays sublimely.

Even Genesis!

(Only reason I mention Genesis in particular is 'cause it sounded totally shite from the 6V6 amp.)

As two units in the chain now have Cosmos pots in this case, the trick here is have one or the other wacked up to full, so it's near as possible working as just like 100k 'earth leak' resistors in parallel across the signal path, and use the other one as the vol control. Best results for the germ amp seems to be, have its own control full up so that the board input is more or less directly coupled to the selector box op-amp buffer, and the selector's Cosmos is then the vol control. Barely much discernible difference 'tother way round though.

I is happy bunny :D
 
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
Post Reply