Rebuild of a dead 1965 AC50/4 by JMI plus an AC50/2 schematic just out of interest.
LETHAL VOLTAGES PRESENT THROUGHOUT THE CHASSIS
1) This old girl came via ebay, dead and very bent. It’s a valve unit with solid state rectifiers made around 1965-66. I owned one of these a long time ago and I remember it being very clean at normal volume levels. I never got to try it flat out because it was bl***y loud and we were playing dinner dances at the time. The Beatles used them briefly, although I guess they had the earlier valve rectifier version and a few more people in the audience. Mine eventually got replaced by a Fender Twin Reverb
2) It's been dropped at some time, probably when the stylish-but-crap plastic handle gave up, and there was some serious damage to the chassis.
Most of the tone pot wiring had been modified (AC50s have a very individual sound and not everyone gets on with them).
Early versions have back-to-front tone controls that cut the bass or treble as they're turned clockwise. Same idea as the AC30.
The speaker connectors had been changed to 1/4” jack sockets and the mains cable was rewired directly to the switch. Later amps had jacks fitted as standard, but this one was originally fitted with XLRs. The good news was that most of it was there and the cabinet hadn’t split.
3) Most amplifiers of this age will be covered with a mixture of dirt and wax. If everything was working I’d normally leave well alone, but this isn’t a particularly rare amp and it was so badly damaged and messed with that it seemed sensible to strip the whole thing out and start again.
The chassis layout was documented before taking it apart and you can see drawings at the bottom of the page. The schematic is a CAD version of the old JMI OS/072 circuit diagram.
4) Straightening the chassis meant taking everything off and doing a bit of careful panel bashing. The vertical section is aluminium so you can’t just hit it with a hammer. Use planed wood blocks above and below the chassis to protect the finish. This one obviously landed on it's back - the transformers will often bend the chassis in the opposite direction when the front hits the ground first.
The steel section is easy to work on with the transformers removed. It was quite badly pitted under the dirt, but came up very well eventually. Wire wool and oil is about as abrasive as you'd want to go on steel or you run the risk of scratches showing up. Aluminium responds well to a soft cloth and metal polish (Brasso).
Heavy corrosion will need fine emery cloth and light oil to get back to sound metal. The oil will keep the dust down but rubber gloves and a face mask are a good idea in case the old plating was cadmium based. Ideally the chassis should be re-plated afterwards although any scratches left will show up clearly.
Some of the protective plating had disappeared on this one and I tried something a bit different and coated the panels in bee’s wax (couldn’t afford replating). So far the results have been good and the smell when the valves warm up is something else.
5) Modern replacement parts have been used where necessary, mainly pots and caps. New-old-stock might be more appropriate for something more valuable or rarer, but this is intended to be a gigging amp and it's got 400volts flying around. I don’t want to be holding the guitar if it goes bang.
(My family just don’t want me to hold a guitar - bit harsh I thought).
The main exceptions were the valve bases, they're all period NOS replacements. All of the electrolytics are new. The original values in the original case sizes are like hen’s teeth to find because modern devices are much smaller and the preferred values are different. It’s possible to open up the old cases and fit the new ones inside if originality is an issue.
Important ones to get right are the 0.1uF caps around the phase splitter. Aerovox 1000v units are ideal and still available on the internet. Everything else is available from TAD in Germany. Expensive but good service. Replacement pots are by Alpha (from TAD).
6) All of the Erie ceramic cased resistors measured near enough within tolerance and were cleaned up and left alone. Likewise the Wima caps. If possible, try not to clean the resistors. All the paint will come off.
The two bias pots were sourced from Surplus Sales of Nebraska. The originals were both open circuit towards one end and what the bias current had been running at is anybody's guess.
The new 16 + 16uF electrolytics were too wide to fit between the tag board and chassis lip and were mounted on made up extension brackets. I’ve seen other chassis where the tag board is fitted further up towards the valves so this may have been an issue at the factory. All the screw holes in the new brackets line up with the existing ones so the chassis is still original.
7) More good news. The Brimistor was in one piece and working. This is a CZ4 negative temperature coefficient (NTC) thermistor made by Brimar and they're impossible to find. It's job is to limit the switch-on current into the reservoir cap.
Update 21st Sept 2012: Gérard Michaud kindly got in touch with a better idea.
"The AC-50 I am presently working on has no Brimistor, but your schematic has it. Was that part there to slow down the B+ ? for 2 minutes? Output tubes like EL34 (and most others) need to be warmed up before B+ is applied, to avoid `cathode stripping`.This why most amps have a standby swich."
plus some advice about the phase splitter capacitors:
"Phase splitter: 1000 v ,.1mf caps: better use 600v caps ,the inductance and the sound quality may vary."
Gérard runs Gérard Michaud Électronique in Quebec, Canada and also contributed some service tips for the Echolette Showstar page.
The CZ4 is a wire ended version of the clip mounted C4. It starts off at about 800ohms at room temperature and then self heats and drops to something less than 100ohms at normal working conditions. "Normal" isn't strictly the right word because the self heating depends on the current drawn and this will vary with bias and signal levels. Later amps or repaired ones may just have a 22R power resistor in place of the Brimistor or no limiting at all. A Welwyn type W22 or W23 would be an easy-to-find replacement.
Max Voltage Factor "k" 1.92
Emax (Volts) 20degC 14.7
Max Operating Current (Amps) 1.25
Resistance at max operating current (Ohms) 5.5
Max Instantaneous Current (Amps) 2.0
Don't worry too much if your chassis has a resistor or direct connection instead. Vox certainly went that route later but I don't know wether their reason was sound, cost or availability.
This thing gets very hot and it's carrying full HT volts so don't be tempted to stick body parts near it. Brimar recommended leaving at least half an inch of free wire at each end before soldering to reduce the heatsink effect of the mounting hardware.
The Brimistor can take several minutes to cool down when the amp's switched off and, if you turn it on again before this, the current limiting won't be so effective. That's one of the reasons your mum told you not to keep turning that old television on and off (well she would have done if you're as old as me).
There seems to be a lot of argument across the internet as to wether the Brimistor has a negative or positive temperature coefficient.
A reasonable approximation of it's characteristics would be:-
R = Ae(b/T)
R = thermistor resistance at temperature T
A = a constant
b = Beta, the material constant
T = Thermistor temperature (K)
Sorry if you've lost the will to live by now.
8) The amp came with a very sad looking pair of Tungsram EL34s, one with the locating lug broken off in the holder. A look around for some modern replacements turned up a very good review of some Russian Svetlana Winged Cs.
£48 later (for a cryo treated pair) and yes they were good, BUT the bl##dy things were a quarter of an inch taller than the old ones and wouldn't fit in the cabinet!!!!!
Not to worry, they'll fit in something else. Anyway, I had another look at the Tungsrams and they turned out to be rebranded Mullards made in Blackburn ..... gold dust. The broken lug was Araldited back on and both valves put on the AVO valve tester. They were both fine.
The only other valve to give any trouble was the front end ECC82 (12AU7). I tried several used and NOS replacements but they were all noisy or microphonic. I gave up and fitted a new Electro Harmonix - very quiet and sounds great.
9) The speaker output panel was interesting. It was factory punched for both the round Amphenol XLRs and the rectangular Cannons. The dents showed that it originally had round ones, so - back to the internet. I had a huge piece of good luck and found a lady in America who had half a dozen original Amphenol 91-855s and who very kindly took the trouble to ship them to the UK. Thanks Cheryl.
The impedance selector link was missing. The back of the connectors on the paxolin panel had been chopped off and the output wire had been hard wired to one side. This was a real pain to sort out and I ended up making a new panel out of several old ones. The new link is a Leak speaker plug with a shorting wire
There were any number of ways that Vox connected mains cables to amps, ranging from four pin XLRs on old AC100s to nailing the cable to the slider board on later AC50s. This one originally had a round connector of some sort in the back panel and had since been modified.
I've got every sympathy with people who get fed up with losing the odd-ball mains cable (and speaker leads come to that) and go for the jack socket / straight to the switch option. However, I've got plenty of spares and the amp looks better without big holes in it. I haven't been able to get an Amphenol shell yet so the new mains connector is a Cannon LNE.
10) The front grill cloth (torn) and the Vox logo (broken) have been replaced with new parts. They look a bit bright at the moment, but original period parts would be far too expensive.
Other things replaced include the odd collection of nuts and bolts the box had collected over the years. Back panel screws should be brass, 2BA x 1", slotted, raised countersunks with cup washers. Best I could do was the same thing in stainless steel. These are still made for classic cars believe it or not. New cage nuts and bolts holding the chassis to the slider board are M5.
The new handle is a genuine old plastic part I've had for years with new metal brackets from a TAD replacement. These are chrome instead of brass coloured, but they match the stainless screws. To fit them, warm up the ends of the handle with a hair dryer and the plastic will soften. Unfortunately it does that while your using it as well.
Carrying an AC50 by the handle .................not a good idea.
11) Speaker wiring.
If anybody wants to refit XLRs, this is how they were wired.
12) So what does it sound like?
Well, with a 335 and a Marshall 2x12, I think it's the dog's dangly bits, but others may disagree. I haven't tried it with single coils yet. Obviously there aren't any effects, so a pedal or two might come in handy.
Should be good for another forty years or so (which is more than you can say for me).
Incidentally, don't expect to get fifty watts output if you're testing with a continuous sine wave. Forty watts rms at clipping would be doing well.
13) AC50 Schematic ( AC50/4 OS/072 )
This is a CAD version of the original JMI OS/072 diagram. The AC50/2 valve rectifier schematic (OS/053) is shown for comparison below.
Left click to enlarge.
Just out of interest, the next schematic is the GZ34 rectifier version, the AC50/2, taken from serial number 01821 (late 64/early 65). Some of the component values and pin numbers vary from the original OS/053 drawing. The shorting links on the low sensitivity input sockets weren't connected to ground as they were on the Vox drawing.
Notice that the phase splitter is completely different to the later model. It still has a BY100 semiconductor rectifier in the bias circuit though.
The NT2 device in the grid bias circuit is a cold cathode voltage stabiliser manufactured in the UK by Hivac. CV2213 is the military spec number.
Thanks to tubecollector.org, r-type.org and radiomuseum.org for the spec.
Max striking voltage = 85v (nominal 80v)
Mean voltage drop across tube at 0.5mA = 60V
Regulating current range = 0.3 to 0.5mA
Max current for continuous operation = 1mA
Max device dissipation = 0.06watts
With only one bias adjustment pot, these chassis, ideally, need a fairly well matched pair of EL34s
BELOW: AC50 GZ34 Schematic ( AC50/2 OS/053 )
NB: The voltages shown in red are specific to this particular chassis. Other amps may vary.
The preamp component layout and wiring. Left click to enlarge.
And the AC50/2 GZ34 version (just for fun). Left click to enlarge.
15) Power amp.
The power section layout and wiring.
16) Top view.
General layout of the chassis.
17) The padded cover was made up specially by Silverstone Amplifier Covers. The padding has proved very effective and a great replacement for the old single ply original (which was missing). It's a very small investment to keep a decent amp in good condition.
Update June 2015. Unfortunately the company seems to have disappeared. Shame.
18) Possible combinations of speakers.
The speaker sockets are wired in parallel inside the cabinet. Any speaker combination much lower than 8ohms, although it probably won't do any harm, wasn't catered for in the original design. Numbers like 8ohms and 16ohms are actually only rough guides. In use, the impedance will vary considerably with frequency, depending on cabinet design and speaker construction.
This is what the impedance of a nominal 16ohm, 18", Foundation Bass cabinet actually looks like.
The occasional accidental short circuit shouldn't be a problem. Gibson, for example, fitted speaker sockets that shorted the output transformer secondary if the speaker cable was pulled out. At the other end of the scale, valve amps aren't too happy with high impedances or open circuits. An AC50 shouldn't be operated without an appropriate speaker or dummy load.
So, if you need to use two 8ohm cabs, wire them in series into one connector and set the link to 16ohms (8 + 8 = 16). Similarly two 4ohm cabs in series into one connector will give 8ohms (4 + 4 = 8).
Both of these options need special leads and aren't shown below. Mixing two cabinets with different impedances is obviously possible but the performance is unpredictable.
Sensible advice part:
Try to operate the amplifier with speakers that match the back panel settings.
Real world part:
It's your amp and you can do whatever you like to it.
19) Layout update.
Thanks to Ken Cumberlidge for sending in the photograph below. It shows the position of C26, the 0.05uF cap across the HT winding of the mains transformer. I've not shown it on issue 3 of the layout drawing so I've updated it on issue 4. C26 is the large brown cap just above the rectifier diodes.