Leak Stereo 30 Plus

A friend of a friend, Ian, got in touch ...

"I've got some Leak kit that could do with servicing. Could this be of interest?"

I like Leak kit.

Yeah.. why not?

He's dropped off a couple of amplifiers (this one, and a Stereo 20), and a pre-amp (a Varislope 2, to match the Stereo 20).

First up , the Leak Stereo 30.

This was Leak's upgrade to the Stereo 30, which was an all germanium transistor affair, this is all Silicon. There's a pair of 2N3055's or equivalent in the output stage, developing around 15 watts per channel into 8 Ohms.

Opening it up , and it's obviously had a life! Lots of sad old electrolytics, in desperate need of replacement. Many leaking.

The main smoothing capacitor, has obviously been replaced in times gone by... The replacement was  somewhat smaller than the clamp would take, so was wrapped in tape and stuffed into a cardboard tube! Even then it wasn't tight, and was wobbling about!

Nasty.

The mains switch has failed, and has been bypassed with a bit of choc-block. 

 

The amplifier is divided into a chassis, which contains the bigger capacitors and mains transformer and output transistors, and 4 plug in PCB's... two pre-amps and two "power amp" boards, although the output transistors are mounted externally.

Here are the power amp modules. The left hand board is recapped and ready to go.. on with the right!

 ... and the pre-amp boards, this time the one on the right is done.

As the amplifier is a single-rail design, the output stage sits at "half-rail" when idle, the speakers are coupled to the output stage via a 1000uF coupling capacitor. It's imperative they are in tip top condition to prevent damage to the amplifier or speakers. They are both physically and electrically leaking.

The first one is removed, and has it's gizzards removed! 

Sadly the fibre top broke up, limiting our options a bit for re-stuffing. I scratched by head for a couple of minutes, and then invented a solution!! 

Ladies and gentlemen (and non-binary persons), may I present ... the Cap-cap (TM)

Empty cap, sans lid, ready for re-stuffing.

New capacitor fitted to Cap-cap (TM) ...

... and the Cap-cap (TM) and cap, fitted to the old cap. I hope that's clear !








Get the model from https://www.thingiverse.com/thing:4860779 . It currently only fits 1" capacitor cans, I may make other sizes as I need.

Very smart.

So, back to our "cardboard tube" capacitor. There seems little point in restuffing it, as it doesn't fit anyway. Back to the 3D printer, and an adaptor is made..

... which snugly holds the new capacitor into the clamp. Neat.

With the capacitor replacement complete, all the pots and switches are cleaned up, and it's time to apply some power and see what happens. 

The two red wires to the top transistors are disconnected and multimeters inserted into the circuit to measure the quiescent current. 

The bias controls, which set the quiescent current are located on the two power amp PCBs.

Well, the bias is all over the place, and isn't stable. The two pots are cleaned up, and ... hang on a minute... 

... that bias pot... why is there solder on it? Someone's obviously attempted to make the contact a bit better by soldering the top spring of the wiper. Nasty. Really nasty.

After replacement of the pots with some nice multi-turn jobs, the bias is set and is nice and steady.

It sounds good, but the balance is miles off, the balance pot needs to be set to the right for it to be central. Also as the balance control is rotated more to the right, the left channel starts coming up again, but horribly distorted. 

The balance control is disconnected and measured, and the resistance values horribly non-linear over the movement on one channel. It's a stereo 20K linear pot. Easily available, and duly replaced.

... not so the broken mains switch section of the volume control. After a fruitless search of the internet for a suitable replacement, a few telephone calls are made, and I manage to secure a new-old-stock leak delta control, which will do nicely.

The amp is now done, and performs wonderfully. Should be good for another 20 years :)

Sony TC-161SD repairs.

Colin phoned

"My mate Susie's got a Sony Amplifier only working on one channel, care to take a look?"

Yeah... why not.

It's a Sony TA-1055. 

It was the world's easiest fix. A noisy tape monitor switch, which, after a few operations put itself right.

Susie also has a Sony TC-161 SD cassette deck. 

It's a top loading deck dating from 1972. 

This one doesn't play or record... Let's have a look...

We need to get in from the bottom. There are 6 screws, some recessed, in the bottom panel. This enables us to remove (after a bit of wiggling) the plastic bottom, and wooden case in one go.

This machine is lacking take up, but rewind and fast-forward work after a fashion, although at a somewhat leisurely pace! The fault's are bound to be mechanical.

The machine is able to detect movement of the tape using a small hall sensor, located on the pulley which is fed from the reels, and eventually leads off to the tape counter. If the hall detector stops seeing tape movement, it engages a solenoid, which stops the mechanism. This is how the autostop mechanism works.

Whilst the belts aren't in great condition, they are working, but are replaced anyway. The small idlers (there are three, one take up, one rewind and one fast-forward) are cleaned up with rubber rejuvenator, but are in remarkable condition. 

Now we have great rewind and fast-forward, but still no take up, which the autostop mechanism detects and stops the deck before our cassette is a mangled mess. The take up idler is not engaging. 

Sadly, it's rather buried under one of the capstans (This is a two capstan deck). You can just see it beneath the large grey capstan flywheel. 

Some disassembly is required...  

Removal of the capstan and flywheel gives us good access. 

The idler swivels into position, and tension is applied by a spring to engage it between the capstan and the take up clutch.. except it can't as the joint in the arm has seized solid. A little application of localised hot air from the SMD hot air soldering station frees it up, and when it's cool, a spot of oil is added. It now moves freely.

With everything reassembled, we now have all functions working... 

... except listening to the output, we've got very weak audio, and occasionally oscillation. Putting a blank tape in the machine and working the record button several times clears the fault. It's a tarnished record/play switch. A quick squirt of servisol helps. OK, we've now got playback but no recording. I can *just* hear audio, but it's very faint. Just as I'm beginning to suspect the record-play switch again, I switch Dolby off. Up comes record audio. Switch it back on and it's a bit intermittent. Another quick squirt of servisol to the dolby switch and all is well. I've also done the limiter, and tape selector switches as well as the record volume controls, just to make sure.

Arduino Audio Compressor.

As part of an upcoming project, I need to compress some audio a bit. 

Now there's plenty of analogue compressor schematics on the web, but a lot require obsolete FETs or odd ball lamps, shining on an LDR (the optical compressor). I fancied a different approach.

Here's the plan, audio comes in to the "top" of a digital pot (also known as an R-DAC), the pot is controlled by the arduino. The wiper of the pot is connected to out audio output. That way, we can use the arduino to effectively wind our pot up and down to control the level of the audio. 

The wiper of the pot also feeds a rectifier, and the resultant level sampled by one of the Arduino's analogue ports, and that value is processed to provide the control signal to the pot.

A schematic is scratched out...

Audio comes in on J1, C3 is a DC blocking cap as the audio is biassed to half of the 5v supply by R1 & R2. The audio then feed the top of our digital pot, U2, on pin 5. Pin 7 of our pot is the bottom of our pot, and is connected to GND. Audio emerges from the wiper of out pot on pin 6, there's another coupling cap, C9, and the audio is then biassed to half our 12V rail by R14 & R16. It feeds a buffer amp, U1B and is output via R15 & C11, and is referenced to gnd by R17.The audio also feed U1A, which has a gain of about 14, and boosts the audio up to a sensible 10 volts pk-pk or so. This is coupled via C8 to our rectifier (D1 and a small filter cap of 10nF. This then drives the ADC of the arduino on pin A0. I haven't really shown the power supply, but it's just a single 12V supply, the 5V being derived from the arduino's on board regulator.

The software, can be found on my github page, as usual  https://github.com/andydoswell/audio-compressor

There are some variables to play with, being Target, Avg, Attack and Decay. Play around with the values and get a feel for it.

One other thing I played with, was feeding U1A from the incoming audio, rather than the devices own output. It was possible to get the software to lose control and end up with no audio coming out, but it could be tamed.. I think this type is called a "feed forward" compressor, and again, worth some experimentation. 

Here it is lashed on a bit of breadboard. It performed surprisingly well, and quiet too.

Heathkit MA-12 Hifi amplifier restoration.

Neil called ...

"Got a load of old 'eathkit gear here , could use the once-over... Care to have a look?"

Yeah... why not?

There's four MA-12 amplifiers, two USC-1 "Stereo control units" (that's a pre-amp to you and I), and an AMF1 tuner ...

The amps are in various states of repair ...

Including one with a valve down to air ...

It's a simple push-pull amp, with an EF86 front end, an ECC83 (12AX7) phase splitter and a pair of EL84's wired ultra-linear. Hmmm , we've seen that before, and looking at the schematic, it's a Mullard 5-10 in disguise..

Off with the bottom ... 

Nice tag-strip construction, and, as these were sold as kits, the assembly looks OK.. some of these can be a disaster area.

Obvious faults... it's full of Hunts capacitors (evict on sight) , and the electrolytics are, well.. past it.

First things first, get that can out, and re-stuff it with nice modern parts ....

The cap is simply removed by gently prying it out with a screwdriver, and saving it.

Next drop the whole can into a container of boiling water for about 15 mins...

Wind a screw into the capacitor's innards ... 

and extract! Probably best wearing some gloves whilst you're doing this bit, I don't know what chemicals are in there ... 

This can is a dual 250uF, 60uF , 350V device. I've got a modern 220uF and a 68uF to go in there, so this is assembled and wired... 

.... and slid back into the can. The cap is sealed on with a little RTV.

The other can is a dual 50+50uF. As it's rubber cap is visible, we want to leave that intact. 

So the bottom bracket is warmed up slightly, and pried off.. 

and the bottom but off with a small pipe cutter... 

It's stuffed with two 47uF caps. They're slightly different shapes so they fit inside the can. 

The base is then sealed up with a splodge of hot-melt. 

The Hunt's are evicted and replaced with some nice polys. 

The two cathode bypass electrolytics (22uF, 25V) are replaced.

Testing shows the output is weak (our valves may be old), and unacceptable distortion...

Attention is turned to the output stage ... 

DC conditions on the two valves should be closely matched... they're not.

The two cathode resistors R21 and R22 have risen significantly in value, one to more that 3 times it's value, they're swapped out. R23 and R24, the screen gird resistors, are also out of spec. Changing this lot puts matters right. 

Now to do the other three! ....