The comprehensive Linn Valhalla repair guide.

John rang...

"Got a few Valhalla power supplies here that are defeating me, can you take a look?"

Yeah, why not.

Now these were the stalwart LP12 PSU once they'd decided that a crystal controlled supply was a good idea. Don't forget boys and girls; Linn ownership is all about upgrades, right?


Be aware all of this power supply is not isolated from mains. Use an isolation transformer during repair and fault finding. Make sure not to inadvertently earth the GND with your scope. You're no longer isolated if you do!
See the red bit in the disclaimer.

It a simple circuit. A 3.768MHz crystal is divided down to get a 50Hz square wave by IC3 (a 4060), and IC5 (4013) . This is switched through some control logic (IC4, another 4013), via the one preset on the board , to a square-wave to sine wave converter formed by IC2 (LM3902), which also provides some gain to drive Q1 & Q2 which drives the output stage formed by Q's 3 & 4. As the output is single ended it's DC component is removed by C18, and supplied to one phase of the motor. A phase shift capacitor is used to drive the other phase.

Most faults can be cleared by changing ALL the electrolytic capacitors, there are 7 in total.

C1,2 and 8 are 47uF 250V axial parts.
C3 is a 220uF 10V axial.
C8 & C15 is a 22uF 50V radial.
C14 is a 1uF 50V radial.

Use quality 105 degree parts. John had already done the caps ...

If that doesn't put it right, check for the presence of 320VDC on the positive end of C1 to GND. This is the HT supply. If that's missing, check the mains fuse (doh!). I've seen the bridge rectifier (BR1) fail a few times. If it's one of those odd looking square or rectangle things a Vishay W10G-E4/51 is an ideal replacement. R1 (47R) is also a frequent flyer!
A short circuit output stage will also cause violent fuse blowing!

Check for the presence of about 9VDC across C3. This is the LT supply. If it's missing, check R2 and R3, both 15K/5W for high or open. Change for a 7W part. They run damn hot. I always try and mount them a little off the board to allow for a bit of airflow. Some older boards can be quite scorched! IC1 can fail giving intermittent or no 9V. It can be replaced by MC3386/MC3346/LM3045/LM3046/LM3144 or an ECG912 (I don't think I've ever even seen one of those!) Z1, a 7.6V 0.25W zener diode, can also go short, open or drift. If you get an unstable or intermittent 9V, change this before IC1. If the 9V is being loaded down by a fault on elsewhere, generally an offending IC, which will reveal itself by getting damn hot! If you're really unlucky (and it has happened) I've seen instances where all the ICs have failed. Other symptoms of a poor 9V rail are inability to set the 85V up, and dim/no LED, even if the supply appears to be working.


Now the supplies are present, check for a 200Hz square wave output on pin 3 of IC3, if that's missing, the crystal or the IC itself may have thrown in the towel.










Square waves present? Good ... Now push the start button. 200Hz Square waves should be present on pin 12 of IC5. If they aren't there suspect the switch itself, or IC 4 or 5.










OK, you should now have a reduced amplitude 50Hz square wave on the end of R14 nearest the pot (there's a 4V DC offset here). If you haven't the pot has gone open (caught me out once!).










Scope the waveform on pin 14 of IC2, it should be a nice healthy 50Hz sine wave. If it's not, replace IC2.











Measure the voltage on pin 10 of IC2. Should be 1.9VDC with respect to GND. If it's not, check the value of R32 (560K). It often goes high, upsetting the bias of the output stage.

Still not working? Check Q1 (BC327, tends to fail open), Q2 (very rarely fails), Q3 and Q4 (either 2SC2501 or BUX84). The BUX84 has a nasty habit of going low gain, and causing issues. Measure it's HFe, should be over 20.

D2 very occasionally fails open, causing the motor to hum or not start. C19 causes the same issue.

The motor is often at fault on high mileage units! Check by substitution. Get a motor from Rega, it's the same and less money! Unlike the Axis, the motor will happily run all day without belt or platter.

Be careful when changing parts as the print quality varies from reasonable to awful. None as bad as the Axis!  There are many slight revisions to the boards, many have slightly different markings, a transistor can be marked Q,VT or TR. IC's U or IC.

Finally adjust the pot for 85V on the negative end of C18, or between the grey and red wires on the motor.

Now sit back, and enjoy some music.

LG GR-L206CPQA fridge freezer repair.

I'm sat in the lounge watching the news....

Ding-dong. Ding-dong. Ding-dong. Ding-dong.

Funny... that's the noise our fridge makes when someone pushes a button on the front panel.

I walk into the kitchen... it stops...

It appears our fridge is possessed (possibly by the curry I stuck in it's freezer compartment last night)

Ding-dong. Nope, it's altering it's fridge temperature settings on it's own...

I'll take a look before organising a priest....

It's an LG of some vintage. It wasn't new when we got it... ah, found the logo under some of the fridge magnets...

Let's get the front panel off... Grasp the panel and gently pull it off...












Unplug the PCB and retire to the workshop with the front...












The PCB is secured to the front panel with two philips screws...










... and removed shows the three tactile switches. One for freezer temp, one for fridge temp, and one switches the ice making thing into different modes.

Sure enough the fridge temp switch reads leaky, at about 40 kilohms. Checking the others show them to be better, but still a bit leaky at ~200 kilohms.




Sadly, I'm out of stock of suitable switches, so I order some. In the meantime, I give the switches a good dose of switch cleaner, which seems to fix it (at least temporarily, whilst the replacement switches arrive)

Another (and probably damned expensive thing to replace) saved from consignment to an expensive de-gassing and recycling scheme, if not landfill.

Lindos LA102 test set repair.

Morning ... coffee ... switch the power to the bench on and contemplate the day's work..

... hang on, the ever-faithful Lindos LA102 Audio measuring test set fails to start up.

It's dead, nothing. Thankfully I have a spare.

A cursory check round shows the mains transformer to have failed open circuit...

Fair enough, it's been in there since the 80's...











A suitable replacement looks like RS 121-3839, and is duly ordered...









Now it physically fits in the hole, we'll need to switch the phases on one of the primary windings, and one of the secondaries.. The original transformer is wired 0-120 0-120, and 0-6 0-6, the replacement is 0-115 115-0 and 0-6 6-0.  I'm not worried about the 5V difference in the primary.

So let's deal with the secondary first ... This is a picture of beneath the transformer.

Cut the tracks as shown.


















And connect two insulated wires to cross the tracks over...

You can now fit the transformer.

Next to deal with the primary. Cut the original link, and carefully remove the trace completely. Note I'm working on a 240V UK unit here... if you're in that funny bit of the world where you use 110 or 115V, you'll need to work it out so both primaries are in parallel.















Then add a link wire in to couple put the windings in series, and in the correct phase...

















... and reconnect the mains supply.


















Switching on, and it's clear something is not right. The transformer is humming and getting warm rapidly!  (It's never easy, is it??)

Turns out the bridge rectifier is short circuit as well... poor thing is cracked too...

It's replaced, and now the unit powers up, but it's still not right, as the power supply is still under some considerable  load...













... and the final piece of the puzzle is the ancient nicad battery is almost short circuit. I disconnect it, as I haven't ever needed it.
















Now, let's get on with some work!

Sharp RG-3915 In-car radio cassette repairs and restoration.

I bought this on eBay..

"Why?"  I hear you ask...













Because I've been spurred on by the warm summer evenings to finish off the green thing lurking in the garage... and it needs some suitable retro sounds. (Sorry if you're still waiting for the final Sony 9-306 YouTube video)

The RG-9305 is an auto-reverse radio cassette, released in 1983 (OK, it's a bit newer than my car, but hey..) It produced a trouser flapping 8 watts per channel (although some have appeared to have escaped.. read on!)

The seller assured me it was "working when removed" ... which I can only assume was sometime in 1985 ;) Thankfully I use the usual eBay translator for such. "Working when removed" means been sat in a damp shed for 30 years+ and the seller can't be ars bothered to test it. (For more eBay description translations check this out)

... and sure enough I'm not disappointed... the cassette deck doesn't work, and the radio is weak and distorted. The radio comes on for a few moments, then the volume fades away rapidly.

Let's tackle the tape first.

Take off the knobs and the securing nuts and remove the faceplate(s) ...













Remove the two screws holding the front panel. It should just pull away.. The remove the four screws to remove the top panel (two on the top in the label, and two on the front). Remove the top panel by gently prising it up with a small screwdriver.








The cassette mechanism is held in place with two screws on the front panel. Disconnect the motor/control plug adjacent to the motor, and the head wiring plug next to the volume pot.

Pop the eject button off (bottom left), and, although it's tight, you should now be able to extract the mechanism complete...

Excellent. Turn it over....















.. and remove the two screws holding down the capstan flywheel retaining plate.

Remove the two offending belts, find suitable replacements in the belt box, and replace them...









Yeah, they're shot ... ;)

Here's an interesting manufacturing technique... real printed components...

It looks like R120, R62 et al are just printed blobs of some carbon type of material, not dissimilar to the "thick film" type of construction used in some TV's back in the day, but those were usually printed on ceramic, not ordinary PCB material. Even the tracks on this side of the board don't appear to be copper, but some form of conductive paint or ink. What if there's been some flexing and vibration going on and there's a microscopic track break? This fills me with dread ... let's see what happens.

On test and the output is low and distorted. There's about 3VDC appearing on the speaker outputs. Not good.

Reverse engineering the circuit shows the output amplifier is coupled to the speakers by two capacitors. They're leaking.











As there's a few others right close to the amplifier, I change the lot...












Ok... things are better, but there's still precious little output, and what there is, is still distorted and thin sounding. Tracing the circuit back from the amplifier, I get to a uPC1228 op amp. It's a dual amp, but in a single in line (SIL) package, and made from unobtainium.. The DC conditions around the amp are odd... There's a DC off-set on the input, which is building over the first second or so the unit is switched on. The audio is coupled in via two capacitors, and loaded by two of those printed resistors I mentioned earlier... and they prove to be open circuit.

An educated guess is made, and two 100K resistors are duly lashed up. One between pin 1 and gnd, and the other between 6 and gnd.

Audio is restored :)









The replacement resistors are properly mounted.



















Now the case has a chunk missing out of it, across the top where the cassette goes..













I used the remaining corner to bend a piece of tinned copper wire to get the outline shape. This is then flipped over, and warmed up with a soldering iron and melted into the plastic case...

It's rubbed down a bit, and covered with capton tape.










Some car body filler is mixed up, and used to fill against the tape. Once the filler has hardened, the tape's removed and the filler trimmed with a craft knife to the rough shape required.











A little more filler is added...













... and finally sanded down to get the final shape.


















A trip to a local paint supplier to get a matching colour ... a quick browse through the racks shows this Pastikote radiator paint looks a damn good match ...
















The front panel is carefully masked off...













... and painted. I'm rather pleased with the colour match and end result...
































The faceplate gets a coat too to get rid of a few scuff marks, and the unit mounted in an under-dash box ready for fitting

and I've got one of these bluetooth FM transmitter devices so I can play music from my phone! (£7.22 on eBay!)

Another saved from landfill :)

Beko DW686 dishwasher repair

It's been busy lately. If you follow my YouTube channel, you'll be waiting for an update on the Sony 9-306. It may be a little while, as other things have got in the way.

Anyway, came downstairs in the morning, ready for a cup of coffee and a bit of breakfast before going to work, and I step in a pool of water... Ugh.

The dishwasher (aka "the magic chipboard") has leaked everywhere, and not washed the plates either!

I got home and needed to tackle the repair. I have an irrational hatred of dishwashers.

I reset it and it pumped the remaining water out. I tried to run a cycle, but it wasn't having it...




I dragged the thing out and removed the front lower panel. It's full of water, and there's bound to be a float switch in there ..











.. (probably behind that central polystyrene float) that stops it from running when the sump is flooded.
















I tipped the machine forward slightly to empty the sump, but it still wasn't running correctly.

I thought perhaps the level switch wasn't working, so it didn't know how much water it had let in (hence the flooding) I removed the side panel and started looking for a pressure type level switch...














... there isn't one...
















...however there's this. It's a small flowmeter in the inlet pipe.

I did a spot of googling, and found some information about a test mode.














Hold these two buttons down and switch on.











After lighting up all the display segments, this is displayed... hit the play/pause (!) button to start the test running....











After going through the motions, Er2 is displayed. Er2 is the error code for "no water", despite the fact that there's definitely water going in, as I can see it gurgling through the pipes ...









OK, suspicion turns to the flowmeter. I decided to remove it and inspect.

Undo the two pipe clips...












Slide the pipes off (easiest to do the bottom one first). Expect a bit of water to leak out...


















And remove the electrical connector....














With the flowmeter removed, I blew through it. I can hear the impeller turning. I can test to see if there's an output by measuring continuity across the device. Those two oval holes provide access, rather than trying to probe the connector.

It seems to be open circuit.

A new one is about £11 eBay. I decided to see if this one could be fixed...



I removed the small green PCB, and it contains a magnetic reed switch. This is simply a small switch in a glass bead which closes when there's a magnet near it. I stick a magnet near it, and it's still open circuit ...







A quick trip to NP Harding provides a slack handful of suitable looking reeds...











One is duly fitted, and tested... it still reads open circuit with or without magnet.... hang on? What I hadn't done, was check the meter was working correctly! It wasn't. It appears one of the probes has gone open circuit! Aggggggghhhh!

Replacement probes fitted, and the flowmeter tests perfectly. I test the reed I've removed ... it tests perfectly .... damn , I've wasted a lot of time ... I'm so cross with myself!

Right, back to the kitchen and re-install the flowmeter.

Run the test again. Same results Er2. No Water. I measure the voltage on one pin of the flowmeter, and there's a solid 12 volts there.... on the other side there's a varying voltage (because of the pulse train) ... OK , that's working....

There's a fault finding flowchart in the service manual (download it here). It says the next step is to check/replace the controller board... It's mounted in the door....

Undo the screws either side of the latch, and the two holding the decor panel on (if fitted, mine's a built in thing) I chose to remove the decor panel completely, as it's front-heavy with it attached)

Another two screws down each side, and you can pop out the plastic top, and gain access to the electronics.

Note the new meter probes! The plug you see removed here has the two purple wires from the flowmeter. Sure enough one is open circuit! I bet myself a nice cup of tea that it's where the wiring goes round the hinge of the door.....

A wire is temporarily fitted to each end...




























and the test mode run again...

and it passes :)










Now to route a new wire in. I thought I may use the old wire to pull through the new one... no chance as it's completely broken... Just long enough to indicate the break is indeed in the hinge... :(













... So the door has to completely come apart. Disconnect both the springs from the door and use something to support it whilst doing this. I found it easiest to lower the spring tension as much as possible with the two plastic adjusters that are just down from the top of the door aperture behind the seal. You'll need to remove the other side panel as well if you haven't already done so.

A word of warning here ... this panel is sharp. Damn sharp. Wear some thick gardening gloves or similar.

And there's the wiring....












The new wire is threaded through (better to start at the top and work down) , soldered at both ends and given a coat of liquid insulation tape (because I've not got any heatshrink at the moment!)









The machine is reassembled, and test mode gets to P3, which has passed the water inlet issue :)











Eventually it's all back together and washing the plates again... for now. What about those other dozen-or-so wires going via the hinge? I'll worry about those another day!

Another saved from land fill (for now!)

... and I won my bet, and have a nice cup of tea ;)