|
Vintage Audio (record players, hi-fi etc) Amplifiers, speakers, gramophones and other audio equipment. |
|
Thread Tools |
4th Nov 2020, 11:52 pm | #1 |
Moderator
Join Date: Feb 2006
Location: Wembley, Middlesex
Posts: 7,224
|
Mullard LP1173
I'm repairing a Decca sound stereo system which uses two LP1173 modules.
It took me a while to work out what was going on with this very odd design. See this thread For the amplifier to work correctly, it has to have speakers connected. For testing purposes, I connected a 100R resistor across the output and found there was approx 4.5V across the resistor. Fitting a 4.7R resistor brought it down to around 220mV. But as the customer can connect and disconnect the speakers at will, could this not damage the amplifier? I'm thinking of fitting a 150R 2W resistor across the speaker terminals, is this advisable? |
5th Nov 2020, 1:18 am | #2 |
Heptode
Join Date: Mar 2011
Location: Dorset, UK.
Posts: 521
|
Re: Mullard LP1173
This unusual configuration was featured for nearly all amplifier designs up to 10W in the Mullard book Transistor Audio And Radio Circuits (first Ed 1969).
A rationale given for this configuration on page 24, second ed. describes "The collector load of the driver transistor is returned to the live end of the loudspeaker so that the signal is applied between base and emitter of the output transistors. The output devices therefore operate in common-emitter, giving maximum amplification." Most of the amplifiers are shown for use in record player, tape recorder or radio situations where the speaker is located inside the same cabinet. I think a careful analysis of instantaneous peak current flow through the transistors when first connecting a speaker load might help determine if this is a really bad idea or not. Rich
__________________
To an optimist a glass is half full; a pessimist half empty; an engineer twice as big as need be! |
5th Nov 2020, 5:19 am | #3 |
Moderator
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,871
|
Re: Mullard LP1173
Q2 is the voltage amplifier stage, the one where all the voltage swing is generated. All after this is simply emitter followers with unity voltage gain. Unusually, it's a PNP device, hanging bat-like from the power rail, rather than the usual arrangement with it living down by ground
The available power from an amplifier has an ultimate limit set by the power supply voltage and the speaker impedance... but this assumes that the amplifier can actually swing to, or close to the rails. If Q2 collector was viased by a simple resistor to ground, the current in the resistor would fall off as the amplifier swung towards ground. This would start to rob the output transistors of base drive current, and the ability to swing low would be curtailed. So, a common trick to help the swing would be to split that collector resistor and drive the midpoint via a capacitor from the output of the amplifier. In this way the amplifier's own output drives the midpoint of the split collector resistor to track the output voltage, so there is always more than enough voltage and current for driving the output in the opposite direction to that which the VAS pulls. The VAS transistor can easily be turned on to a far lower impedance than its collector resistor would have to be, so swing in this direction isn't a problem. So this is what's called a bootstrapped load for the VAS, voltage amplifier stage. What Mullard have done with this design is to save one whole capacitor and one whole resistor. The collector bias resistor for the VAS goes to the speaker output rather than ground. The same bootstrap effect happens without needing a bootstrap capacitor, and only a single resistor is needed. It makes the circuit more complicated, it puts DC in the speaker voice coil. Speaker disconnection transients are routed to the VAS stage which is a good deal more delicate than the output devices, and it is already working hard on the voltage front. Bootstrap arrangements can come unstuck on transients in the programme material that try to clip the amplifier. This sort of amplifierusually gets stuck for a while and turns a short transient into a long hang-up. So what you have there is inelegant and risky. Bootstrapping is common in older designs, and in not so high-fi stuff. If you have to fix one that already exists, then , yes, you have to have a load resistor or speaker present to get a channel to work. Ideally a resistor equal to the nominal speaker impedance. It becomes a problem if you want to modify something to have a headphone output... you need to provide a load resistor. This cheapened bootstrap is why the VAS stage is PNP, so that its collector bias comes from ground. Back in the day PNP transistors were slower and less robust than NPN, so it got another disadvantage. Those amplifier designes were aimed at showcasing Philips' semiconductors to set manufacturers. Saving a couple of pennies on a resistor and a capacitor were a big deal to them. It became a pure increment in profit for every box they shipped. No, it's not a good idea. David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done |
5th Nov 2020, 7:41 am | #4 |
Octode
Join Date: Sep 2017
Location: Bristol, UK.
Posts: 1,654
|
Re: Mullard LP1173
David - what an incredibly erudite and clear explanation. Thank you!
Mike |
5th Nov 2020, 12:32 pm | #5 |
Moderator
Join Date: Feb 2006
Location: Wembley, Middlesex
Posts: 7,224
|
Re: Mullard LP1173
What's not a good idea? The whole amplifier design, or me putting a 150R across the speaker terminals?
|
5th Nov 2020, 4:28 pm | #6 |
Moderator
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,871
|
Re: Mullard LP1173
Just the design of the thing. It adds vulnerability and it relies on a DC path that most people would not expect to be needed. If I wanted to build a small audio amplifier, I'd avoid any designs with this trick.
You'll have to add a resistor to get it to run to be able to test it for the purpose of repairing a module. Too big a resistor value will decrease the quiescent current in the VAS and thereby rob the output stage of drive in one direction. So if you run into distortion at higher signal levels, the value may need dropping. David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done |
5th Nov 2020, 6:09 pm | #7 |
Guest
Posts: n/a
|
Re: Mullard LP1173
The only thing I can see as a disadvantage is if you connect the 'speaker while it is on. Hardly likely in a "one box" thing and would only give a thump in the 'speaker anyway just like any cap coupled amp without a bleed resistor. I like it.
|
6th Nov 2020, 1:26 am | #8 |
Moderator
Join Date: Feb 2006
Location: Wembley, Middlesex
Posts: 7,224
|
Re: Mullard LP1173
The Decca has external speakers, connected by DIN plugs at the back.
|
9th Nov 2020, 1:39 am | #9 |
Moderator
Join Date: Feb 2006
Location: Wembley, Middlesex
Posts: 7,224
|
Re: Mullard LP1173
An update: I have not been able to successfully repair this amplifier. I originally replaced all four transistors and resistors R123, 4, 5 & 6 using new devices from Cricklewood electronics.
It should have worked but I didn't realise about the biasing requiring the speakers connected. I've now connected a 4.7R resistor to act as a load, but couldn't get the thing to work correctly. The AC128's had a different can from the usual type supplied in that it looks like an elongated BC107 with a tab as the emitter not a dot as the collector. I swapped it for a BC640 and got nearly the right result but its not good enough, feeding a sine wave into it, its very distorted, much more so on the 2nd amplifier. I could order some NOS AC128's from another supplier but am wondering if its worth it. Or I could rebuild the amplifier completely using a pair of LM3886's. I have suitable boards from the Beocentre 2002 back in March, fortunately the power supply is a -ve earth so all I have to do is build a suitable power supply for the mid voltage. (I'll show you a diagram in another post) Below is the circuit for the LP1173 |
9th Nov 2020, 8:37 am | #10 |
Nonode
Join Date: Jun 2016
Location: Bognor Regis, West Sussex, UK.
Posts: 2,296
|
Re: Mullard LP1173
Hello Michael,
I have a pair of NOS AC128s with the the dot as collector you can have FOC. PM me your address and I will pop them in the post. Peter |
1st Dec 2020, 7:13 pm | #11 |
Moderator
Join Date: Feb 2006
Location: Wembley, Middlesex
Posts: 7,224
|
Re: Mullard LP1173
First of all a big thank you to Peter for letting me have the two AC128's.
I've only just got round to fitting them. On one of them, it appears to work normally in so far as the mid voltage is around ½VCC but the other, although it starts ok, the mid voltage is ½VCC then after around 15 seconds, switches to around 25 V. (VCC is 28V) On that amplifier module the AC128 had tin whiskers which I got rid off by using a nearly exhausted PP3 battery. All other transistors are new, so either I've got a duff AC128 or there is something else going on. For testing purposes, the modules are out of the player being fed with an external power supply and and a 4.7R resistor to simulate a speaker. |
1st Dec 2020, 9:56 pm | #12 |
Nonode
Join Date: Jun 2016
Location: Bognor Regis, West Sussex, UK.
Posts: 2,296
|
Re: Mullard LP1173
Do you want another one to try, I have still got a couple.
Peter |
1st Dec 2020, 10:52 pm | #13 |
Moderator
Join Date: Feb 2006
Location: Wembley, Middlesex
Posts: 7,224
|
Re: Mullard LP1173
Thanks Peter, I’ve ordered a couple from EBay, will have another go at the weekend.
|
3rd Dec 2020, 3:44 pm | #14 |
Tetrode
Join Date: May 2014
Location: Bridgnorth, Shropshire, UK.
Posts: 51
|
Re: Mullard LP1173
Decca 603 system in common with all Decca equipment
using the LP1173 module used 8 ohm speakers ie L100 etc using 4 ohm dummy loads not a good idea. Regards Robin |
18th Dec 2020, 5:51 pm | #15 |
Moderator
Join Date: Feb 2006
Location: Wembley, Middlesex
Posts: 7,224
|
Re: Mullard LP1173
An update:
I've repaired it and it's on test as I write. The original AC128's were mounted in a special package which acted as a heatsink when bolted to the main heatsink. As this could not be used, the transistors were getting too hot. I used some heatsinks from a Dynatron radiogram chassis which were bolted to the main heatsink. It now runs reasonably cool. |