Translate

Saturday 30 April 2016

1971 GEC Colour 2103 TV repairs. 2040 Chassis

My colleague Nathan came into work. He's just moved house.

"There's an old telly in the loft. Do you want it?"

Yeah, why not....

I take the set home, much to the delight of Mrs. Doz











The cabinet has seen much better days, and it looks like it's been damp at some stage in it's life...











No model number, as the sticker's fallen off, but some research shows it to be a 2103, dating from 1971. It's a single standard (625 line PAL) colour set.









... it looks to be complete...












... and un-molested...












Removal of the line output stage screening can, shows the anode lead to the line output valve's insulation has failed, leading to a bit of a burn-up











... and there's a resistor on the side of the line output transformer which falls into pieces as soon as I touch it. This is replaced, and the wiring made good.









The CRT is tested...













... and doesn't look good initially, but after a couple of hours wait, the meters rise into the green as the tube wakes from it's long sleep!











Some mains is gently applied to the set via the variac... nothing. The mains fuse is open circuit, so a new one is fitted... and as the valves warm up....

... a blurry green raster appears ...













.... and after some tweaking, some rugby players appear. The convergence is awful, and the picture isn't in colour, but it proves that the set is viable...









The tuner is seized solid as always, so it's removed...


... and cleaned up...












Being careful not to stress those rusted in plastic nuts!












The tuner is refitted once it's all moving freely, and I start fault finding on the lack of colour.

I prevent the colour from being removed from the picture (for when black and white programmes were broadcast) by disabling a circuit called the colour killer. There's still no colour, so it's likely the reference oscillator isn't running...  Just as I'm about to start looking at that fault, the sound and picture disappear! The set is still operating, but there's nothing coming out of the IF (Intermediate frequency) stages after the tuner. Damn.

There are 3 stages of combined IF amplifier. These amplify and filter both the sound and vision the signals together, coming from the tuner. The signals are then split off, and amplified further, the audio being separated and decoded from the video. As both sound and picture went off together, I suspect the fault lies in the combined amplifiers.... I check all three stages, all appears well. Perhaps the tuner has failed? A super chap called Gary has a spare, but it makes no difference....

After some further checks, largely out of desperation, I check through the vision amplfier, and find there's an emitter follower amplifier, and the transistor has failed. It's a BF194 HF bi-polar transistor. A replacement is fitted, and pictures are restored! Still no sound. I look at the first sound IF, and , sure enough there's another failed transistor. A BF194... still no sound!... I check the ratio detector, which is used to demodulate the FM sound, and.... it's another failed BF194! Sound a pictures restored, I look back to the colour decoder.

I check the reference oscillator, and it's running fine, producing a locked 4.43MHz signal. Good. I check through the stages, and find 3 faults. Two of the faults are transistors in the R-Y (Red minus lumenance) and B-Y (blue minus lumenance) ... guess what .... BF194's. Also the colour bust transistor is open circuit... you've guessed this next bit ... it's a BF194...

There have been questions raised on various forums in the past about the reliability of BFxxx and BCxxx lockfit type transistors. It's quite possible that that un-reliability (they are 40+ years old!) is what's been at work here. It's possible that flash over when the insulation failed on the anode lead in the line stage caused a spike which saw off the transistors, but the BF197's and BF196's in the IF stage survived, as did other transistors in the decoder...

After replacing this little lot, and a couple of suspect and corroded looking capacitors, things are shaping up. There's some hum evident on the power supply, and I find a very sad looking cap in the power supply. It's got a nasty bulge in it... It's changed out.







I replace the line output valve and boost rectifier (PL519 and PY500) and the width is better, set up the convergence on the tube, so the three colour guns are lining up correctly, and I'm rewarded with reasonable results.











Note quite sure what I'm going to do about the tatty cabinet though...

The guilty parties....














Saturday 23 April 2016

Little microphone preamp.

Now, I've not been very happy with the sound on my videos... you may have noticed...

What's needed is an external microphone, maybe a tie-clip or a lapel mic.

Now examining my new camera, it doesn't have a mic input. Damn... I google a few things, and, although there is a guy on a forum who's added a socket in to the camera, I didn't fancy taking my new camera apart. Other people with the same camera seem to prefer to record the sound separately, on a zoom recorder, or similar, and line it up in the video editing software. Good plan.

Now years ago I bought an Acer MP3-330 MP3 player, fitted with a mechanical 10GB hard disk... Now that has a record facility, but the in-built microphone isn't up to much. It does have a line-in socket, however. So we'll need to make a microphone pre-amplifier.

Left over from the Arduino beat detection and lighting controller project I had a number of small electret condenser microphones. One of these would work nicely as a discreet lapel mic.


The circuit is a simple op-amp, driven from a single rail 9V battery. R1 is used to supply the bias voltage to the microphone. It should be around 7.5 volts. C2 is used to block this DC. R2 and R3 put the incoming audio from the mic onto a half-rail supply, which is amplified by the op-amp, having it's gain set by R4 and R5 in the negative feedback loop. C2 gives us some DC stability, but ensuring that the DC gain of the amp is unity, otherwise any offset would be amplified also. C3 removes the amplified audio from the half-rail DC on the output. R5 is a level control. A snazzy blue led is driven at low current from the switched 9V supply.

I make the circuit up on a very small piece of veroboard and mount it, along with a couple of sockets for input and output into a small box, It works well.


I mount the microphone on a long length of screened cable, and enclose wiring side of it in heatshrink. A crocodile clip makes an excellent clip for it.











The finished unit. Watch this space for a demo!

Saturday 16 April 2016

Pye T19D radio from 1949, repairs and renovations.

A fine fellow called Derek called...

"I've got this Pye radio I've bought, and it doesn't work, it's totally dead.. Can you take a look at it?"

Why not ....

It's a very nice looking Pye T 19 D....

A cursory inspection shows a very clean chassis, but the usual rake of the dreaded Hunts capacitors, both wax type and electrolytics. Nothing to stop it powering up though...



.... so we power it up. Nothing. Not a sausage.

A quick check round with the meter show the mains switch is open circuit. The mains switch is unusual; it's mounted on the tone control/ mode switch. This switch is multi-funtion, and switches the set on, and provides a switched tone control, and selection of the "gram" input, for connection to your external gramophone...

Removal and inspection of the switch show it's stuck in the off position. It's a bit of a tin-pot affair, being paxolin construction and pressed together in a steel case. I managed to get it apart, and un-stick it, but it's not reliable and starts arcing over. Drat. A solution will need to be found....








The tone control/ mode switch without the mains switch sat on the end...














In the meantime, I hook up the main smoothing electrolytic capacitor to the Dreadnaught MK87 capacitor reformer, in the vain hope of it being OK...












...it is nicely dated APR 1949, but I don't hold much hope, the poor thing has been in there for 67 years....










 I start to re-cap the chassis, evicting the evil Hunts...

There was also some Sprague 1000v caps that read leaky; they had to go too to ensure reliable operation...















... and when this is complete, I check out the mains smoothing electrolytic, which has been gently reforming for a couple of days. It's OK! I'm amazed. 67 years in, and it's giving acceptable leakage, a reasonable ESR, and capacitance! ... I short out the mains switch and gently apply some power via the variac... and the set crackles into life! It's quickly apparent that's something's amiss with the tuning drive, which slips badly ... perhaps a re-string of the tuning drive is required? ....

Nope, the tuning capacitor is moving on it's rubber mounts as I move the tuning control...

I slacken off the bearings and give them a little lubrication.. Bingo! Smooth tuning and no sign of slippage on the drive string.. Good news!








The volume control appears to be open circuit for most of it's travel. It works at about 90%, but nothing below...

So, to re-cap (pun intended)... the remaining faults are:

An unserviceable on/off switch and volume control...

The volume control is a 1 megohm log pot. No problem there. These are easily available.
The mains switch? Made from purest unobtainium., no chance...

So what about getting a 1 meg log pot, with a switch, and have the on/ off switch on the pot? Seems a reasonable compromise.... but 1 meg LOG pots with switches are also unobtainium...

I go to the local Maplin, who stock a 1 meg linear pot with on/off switch.. and fit it. It's horrible as a volume control... the control is all "up one end" , as you'd expect.









Now a "log" pot isn't really a log at all, but an approximation of log, by damping one end of it with a resistor; so, after a bit of experimentation, I end up with a 47K resistor at the earthy end of the pot. It works well.

I did contemplate mounting the pot on the end of the tone/mode switch, and not bothering with the pot bit, but the mechanical resistance in the pot made the switch unusable, and worse, prone to arcing if it hadn't exactly sat in it's ident.

So this set now has a volume/on/off control. Not ideal, but it works nicely...

Here's the completed set in action...


Saturday 9 April 2016

Junque Box Boutique Amplifier.

After repairing Norman's Rickenbacker amplifier, he came to me and said "Andy, you can't keep repairing guitar amplifiers if you don't own a guitar"

Fair point.

So he very generously provided me with such.

Er ...

... yeah ...



So, after repairing the recent Marshall amp, I managed to wrangle the opening bar or so of "Matchstick Men" by Status Quo out of it, then Steve came along and collected the Marshall, much to the relief of the neighbours.






What's needed here is an amp to call my own. Shall I be going out and purchasing such? No. Shall I be looking round in the workshop to see what we've got kicking about that I can cobble together at minimal cost? Read on ....

Whilst I'd love to make something really big and loud, purely out of engineering interest, I'm sure the neighbours wouldn't.

So, a few bits and pieces were gathered together. I managed to score some unusual output valves for not a lot of cash from eBay seller seemoreitems. Some wire-ended N78 pentodes. I had a couple of mains transformers that I could lash up to get heater and HT from, and a rather nasty 8" 6 ohm Sansui "Woofer" from the 80's....



I was going to need to order a few things, namely a couple of chassis boxes, and an output transformer. I gave Philip at Bluebell Audio a ring, and bought a Hammond chassis and a 125D Push-pull output transformer.



So, gather the items together and see what transpires...


First up, the power supply. I orginally toyed with the idea of using two 6V transformers back-to back, but this was going to end up being huge. I found a nice 20VA 6V transformer, and a 50VA isolating transformer.. the junk box provided a bridge rectifier, and a couple of caps, all salvaged from a duff switched mode computer PSU



 Top side of the completed power supply....












And the rectifier and filter...

Note the tags on the top of those transformers have mains on one side, and our un-rectified HT on the other, we'll need to make a cover for that later...







Smoke test gives around 280V DC off load. Let's push on with the amp itself....

I'm going to start with the output stage and work my way back to the input, making it up as I go along!

First off is to look at the N78 datasheet, which can be found on the wonderful R-type.org website.

So, a pair in push-pull, class AB1, with 240V on the anodes (allowing for a little sag from our cheap PSU), should give us a nice healthy 5 watts or so, with a 9 kilohm anode to anode load. That's quite a lot for a small bottle, no wonder the N78 has a reputation for running hot.

So, how to mount our wire-ended valves?


I did contemplate mounting in a stainless steel cable gland, but it's a bit ugly, and I'm concerned the valve won't be able to radiate enough heat. I ended up drilling out some B7G bases, and feeding the wires though the holes and soldering them to the tags. Perfect. Be careful with the wires, if one breaks off, you've had it.





I mounted the output transformer, and started marking out for the valve bases.

I'll cut three for now, one each for the N78's and a B9A base for an ECC83 (or something) for the phase splitter.







Holes cut and tapped ....



Valves mounted....














Heaters wired up using twisted cable, it helps to keep the hum out. .... so it must be time for ....
The arty valve shot ...

One side of the ECC83 (12AX7) will perform phase-splitter duties, whilst the other will be our pre-amp.









After I got the phase splitter operating, I added another ECC83, this time as a cathode-follower to drive a tone-stack. I shamelessly pinched the design for the tone stack from Marshall, who stole it from Fender...

The other section is another gain stage...







I added some pots, a mains switch and the input jack... eventually it's my intention to mount the amp upside down, with the valves hanging down, so the layout seems wrong. Controls are (from left to right in this picture) "Drive", Bass , Mid , Treble, "Colour" and Gain.






Testing at this stage revealed it needed more "oomph" for that overdrive rock sound... Let's add another valve! ..  another ECC83.

Now things were sounding good... but left me with a problem, I had an unused triode section.

I thought about adding yet another gain stage, but that would just be silly, and probably noisy too... so I set about thinking about a tremolo.



I looked around the valve wizard website, which is a fantastic resource, and cobbled together a tremolo oscillator... I was hoping to couple the cathode of the oscillator to the cathode of the input gain stage to produce the effect. It didn't work...

What I needed was a method to modulate the volume... I had a cup of tea and a think. What about a small FET?



I had a look through the FET drawer, and selected a likely candidate,a BS107 in this case, but a 2N7000 would do nicely.) Connecting this temporarily to the volume pot worked well enough. I added depth and speed pots to the rear, as well as a foot switch plug, to switch the tremolo on and off.






Tidied up the wiring, and added a pink LED (because blue is sooo last year).













Now to tidy the workshop, and think about a cabinet....

Incidentally, this picture was taken with only the heater supply on, to show the valves lit up, and the LED working. Never, ever run a valve amp without a speaker or dummy load connected.






Here's the schematic...



Let's have a little walk through...

The signal enters at J1. C1 is used to block DC, as the grid of the valve is around 1.5V above ground, to set the bias of the valve. R1 is the grid leak, and also sets the input impedance. R4 is the grid stopper resistor, and, in conjunction with the miller capacitance of the valve and C4, stops the thing trying to oscillate or pick up radio. The signal is amplified by V1A, and emerges at the anode (plate if you're in the US). Anode load is provided by R8. The signal is coupled by C6 to the volume control, R10. The signal is then coupled to the drain of Q1, which is the tremolo modulator FET (more on this later..). The signal then passes via another coupling cap, C17 to another gain stage, V2A, in the same manner as the first. The output of the anode at V2A is directly coupled to the grid of the cathode follower V2A, which has no gain, but has a low output impedance to drive our tone stack. The tone stack consists of three filters, created by C11,12 & 13, R18 and the pots, which are used to control the amount of attenuation of each filter. The output of the tone stack is connected to the drive pot, coupled to V3A in a similar manner to before. The cathode of V1A has a pot connected, so we can change the bias of the valve slightly. This has the effect of altering any clipping that may be occurring, and, in a guitar amp, we want a bit of distortion! If the tube is biassed "cold" i.e. less current flowing (larger value of cathode resistor) the distortion has a "bluesy" sound to it, if the tube is biassed "hot" , it's more of a "rock" sound. The effect is subtle. The output at the anode of V3A is coupled by C17 to V3B, which forms the phase splitter or phase inverter. The grid of V3A is fixed at ~53V by the bias network, formed by R29 & R30. An inverted signal appears at the anode, and a non-inverted signal at the cathode. Note the anode and cathode resistors are the same value, ensuring balance (as long as the load is symmetrical). The inverted, and non-inverted signals are passed by their respective coupling capacitors, C18 & C19 to the grid of the N78 output valves. R33 (R34) provide grid leak, and R35 (R36) are the grid stoppers. Anode load is provided by the output transformer. R44 & R45 are feeding the screens. This resistor stops the screen grid from exceeding it's maximum dissipation during overdrive conditions. V1B forms a phase shift oscillator, and operates from about 2Hz to 10Hz, speed being controlled by R2. Each capacitor, which it's resistor, provides a phase shift of 60 degrees, and the inverted signal from the anode is inverted a further 180 degrees, forming the oscillator. The output from the anode is also coupled by a large value capacitor, C7, to R12 "depth" control which is used to feed the gate of the modulator FET, Q1. via the potential divider R13 & R14. The source of the fet is connected to ground, and the drain to our audio line, between C9 and C10. When the FET is biassed on, the audio is coupled to ground via the FET. Shorting the gate to ground switches the FET off, and switches off the tremolo effect.

How does it sound? Great... let's rock ...

.. now to learn to play the guitar !






















Wednesday 6 April 2016

Mobius HD Actioncam and RFI.

Now, being a gadget lover that I am, I decided to upgrade the rather old dashcam in my car to something a little more "HD". After doing a bit of research, and seeing the footage of a colleagues camera, I decided to get a Mobius 1080P Actioncam, purchased from a reputable source (beware the fakes). Good choice. Small unit, great pictures, sound is OK... except....

... it ruined my radio reception.





Now I have a newish car, and it's fitted with a DAB radio. After fitting the camera to my windscreen, I noticed DAB reception of my favourite station, Planet Rock, became patchy. Disconnecting the camera solved the issue.... hmmmm ... RFI ... Radio Frequency Interference.

So what was going on? I broke out my ancient spectrum analyser, and, equipped with a simple whip antenna, took some measurements....

Here's the DAB reception, centred on 220MHz, with the Mobius disconnected. The two humps are the two DAB Mux's (multiplexes) I can receive at home...










... and here's the same shot with the camera connected... yuck. I suspect the big peak to the right of the mux is effecting the AGC of the radio, and there's a lot of hash underneath it too, upsetting the signal to noise ratio... boo!








I tried a simple method of preventing RFI radiating from the cable of the device. I wrapped as many turns of the USB power lead round a random (ex-computer power supply) ferrite ring.









Brilliant....














... and after some field tests, reception restored!


Saturday 2 April 2016

Arduino Audio Wattmeter.

Many, many moons ago I needed something to measure audio power in a hurry. So I lashed up a quick and dirty audio power meter.

It was nasty....

... very nasty, but did the job at the time.

It was simply a switchable 4 or 8 ohm load (actually 4.7 or 8 ohms)  which was rectified, and used to drive two meters, each measuring full scale at 10 watts. The fan was driven from rectified audio. Most unpleasant.




Anyway, the load was good, as was the fan, but I needed some sort of better instrument, so a plan was dreamt up. Measure the peak voltage across the load, do a bit of maths and get the results displayed on a 1602 display, driven by an arduino. Great ....


So audio comes into the load (R1 & R2), and is rectified by a bridge. There's a small bit of smoothing on the output, so we can get a steady-ish DC. This is divided by a voltage divider, R3 and R4 and the resultant is fed to the arduino's A-D converter. Couple of things to note here, the load is switched, either 8.0 Ohms, or 4.7 Ohms, by S1. This also lets the arduino know which load is selected. S2 is a "BTL safety switch". It completely isolates the left and right channel. It also stops the arduino reading the right channel. If you're testing a BTL(bridge tied load) amp, or any type of amp which may not take kindly to commoning up of the negative poles of the speaker, best click this across. It also grounds the second analogue input to prevent spurious readings. A PWM signal is output from D6, and is used to drive a FET to run a fan. I chose a stupidly large FET here, you can pick something more sensible. I just happen to have a bucket of them. 

This will only give accurate results when feeding in a sine wave.

That's it really....

 Dummy load salvaged from the old unit...

Front panel coming together.

... sizing it all up....

... starting the wiring...

Looking good until I noticed that I'd mistakenly picked up a 16x1 display instead of a 16x2!... and the 16x2 requires a slightly smaller opening. Drat! ... some hot melt later will have to do (and it was all going so well ...)

Rectifier board and PWM FET

Rectifier was a little overkill! Still I had two identical, salvaged from a switched mode PSU.











Setting up ready for calibration...

Apply 28.28VDC accross the outputs of the two bridge rectifiers, and with the unit set to 8 Ohms, adjust each pot until the display reads 50.0W. Easy... (Why 28.28V ? 20 VRMS into 8ohms is 50Watts. 20*1.414 = 28.28Vpk... see here for more)





All calibrated and ready for action!
Top line on the display reads left channel, bottom reads right. The three readings are instantaneous power, average and peak.











And last, but by no means least... the code....