Laney Klipp 100W Repair and Design Study – Part 1

Laney Klipp 100W Repair and Design Study – Part 1

DISCLAIMER:    READ THIS!!!

 

http://en.wikipedia.org/wiki/Electric_shock

There is a big effect difference between voltage levels that break through the skin conductivity – which happens to be in the 400 to 500V +DC range of our valve amp circuits! BE CAREFUL! Drain and check WITH A METER the PT secondary capacitors when the circuit is off before touching components or working on it – with the mains UNPLUGGED. If the Mains is plugged in, the plug socket or wires and switch contacts are still LIVE, with the amp switch in the OFF position.

If you are testing the circuit after a change, make sure all you test leads, probes and croc clips are well insulated from your fingers and chassis contact points (chassis edges can be sharp too). I always connect the croc clips and meter and double check. I then plug in the fused mains side. I use the mains toggle switch to power the secondary with the indicator light connected so I can SEE power also, as well as checking the on/off position of the switch BEFORE I plug in the mains. I switch OFF and unplug the mains, checking the meter voltage level across the power transformer secondary filter caps has dropped to 0V, before I move any croc clips to a new circuit point for another measurement.

For the sake of a few extra seconds, why risk safety by not switching off and watching the circuit drain? Most valve amps take only seconds for this to happen unless the valves are out, when it can then take minutes. Anyone can slip or drop a probe on a live circuit or be distracted when checking something. Why risk it? Get in the habit of checking the meter and mains plug before you touch the amp.

A 240V AC shock will certainly help you not make a mistake again if you ever forget any of this – if you are lucky enough to get away with it! My last one left my right arm aching for an hour!

A 500V DC shock may mean you don’t get to worry about any of this ever again…but now you know why if you read the Wiki shock link.

———————————————————————————————————————————–

This Post was a great help in understanding some unusual aspects of this amp before checking it physically too much, to know what to expect where to some degree, and as a general theory recap just by studying the schematic first. It is tidier than the JMP was and now less scary, even with a 600V+ HT, as I am aware of some of the issues that can go with that (screen voltage limits on newer EL34s for example) from stumbling across relevant info on forums etc. whilst looking for other stuff.

As always, Knowledge Dispels Fear. Don’t confuse this with taking an amp for granted though and get complacent – ever! Never work on an amp when tired or with distractions around or many possible interruptions. Common sense at all times and if in doubt, stop and research it.

I have the Variac also to bring this amp up slowly anyway, which is recommended for any old amp with older or original components – electrolytic capacitors being the biggest risk of failure and further damage.

The schematic study brought to light some odd wiring like the PT sec fuse being attached direct to Mains Earth, which is also NOT connected direct to the chassis in my amp. This appears to be a change made from laziness* when the original mains cable was changed, not bothering to attach to the chassis but direct to the fuse holder.

*Nope – One Marshall JMP schematic I have also wires the secondary fuse this way, from ground.

*********************************************************************************

Just to re-list my rough method for checking an unknown amp, as I haven’t done a repair for a while before I look at Klipp related stuff:

*******************************************************************************

1 Test the amps valves in the Ashton kit amp/valve tester/known working amp.

It can be good news to have one failure, as this is probably the only reason for a non techie owner selling as a spare/repair bargain.

2 Check the plug Earth to chassis connection – 0.5 ohms or less – and all fuses present and correct type and value – SAFETY FIRST!!

3 Check all the caps with the scope component tester – get an ellipse for all but the smallest values –or use a cap tester.

Always check the big filter electrolytic caps with a cap tester for actual values and any connection to the chassis (ground) is 0.5 ohms or less. If values are way off, then possibly they have failed. Look for bubbles/leaks and general condition. (All this won’t guarantee against an old cap blowing under full HT of course.)

4 Check the rectifier diodes with DDM to check 0.5 ohm forward conduction and open circuit reverse bias.

5 Maybe check all/some resistors and caps with scope/DDM and cap tester depending on condition of amp overall or burned/tired looking specimens while referring to schematic for those in parallel that may give false impression readings.

Always check the continuity of all inductors – PT, OT and choke. Check for an ellipse on the scope component tester for low resistance windings. The heaters may not show well so may appear shorted but are not.

6 Check all switch function using DDM.

7 If happy enough overall, use a Variac (for old amps) to mains test with the valves removed, checking diode A+ voltage (e.g. 510V DC), and ripple (e.g 7Vpp) on scope which shows rectification is occurring.

8 If no smoke, fires, bangs etc. power off, timing the discharge time in seconds (I can hear my wall clock doing this while writing down a voltage every 5 secs if I want to do a graph in Excel later) so you know how long it takes for amp to become safe to touch (e.g 510V-20V in 20 seconds)

9 Replace valves and redo step 8 with all knobs at zero and a speaker connected.

10 Increase volume gently until mains hum/hiss is heard. Poweroff.

11 Setup FLS test signal at 200mV on scope for a powerful amp. Plug FLS into an input, make sure volumes are down. Connect test probe to speaker out jack for testing end to end test signal distortion.

13 Power up again, increase volume of input and main volume gently until sine wave is heard. Check input sine shape against output shape for symmetry, and phase relationship (usually in phase unless NFB used). Poweroff.

14 Do any further tests you wish like Freq Resp, LTP distortion stage or test with a guitar for tone and all control functions – crackly pots, presence, bass, mid, volume etc, all input and output jacks, reverb tank (if connected).

15 Re-assemble, test and take to studio for a mega blast through a 100W cab, with rest of your kit, and video it!

*********************************************************************************

Laney Klipp Amp

http://www.laney.co.uk/history/laney-article

http://www.namm.org/library/oral-history/lyndon-laney

Initial serial number research:

“Serial number is 33X872. I believe those klipps were made between 1970-75. So to draw a conclusion… maybe (correct me, if mistaken) the last two digits indicates the year…1972!

Fascinating theory on dating head! The serial Number on mine is 33X617, so by your reconing it was made in June 1917! seriously though, there is a paper quality control label stuck onto the chassis of my head. It has various boxes marked Batch No, Box No, Box by, Assembly, Chassis, Connect and Test. The only boxes that have been completed are Batch No and Assembly. There is a squiggle of a signature in the Assembly box, and the batch number is 7/74. Could this mean July 1974?”

“I’ve just had a message from a very helpful person telling me that the serial numbers on the transformers date them to 1971”.

Partridge Transformers:

Dating Partridge transformers…it’ll cost you though.

by Goku on Tue Aug 14, 2001 1:55 pm

Thought I would pass on an email I got regarding dating of Partridge transformers from the <A HREF=”http://www.transformers.co.uk/frameset.htm” TARGET=_blank>manufacturer</A>.<P>Yes, we are the manufacturers of Partridge Transformers, and yes, there is <BR>a dating system used. However, this is based on our internal contract <BR>reference so it is not a system that we could practically pass on to you. We can search for the date using our contract records but there would be a <BR>charge of 50 pounds sterling for the first item, and an extra 15 pounds for <BR>each additional item handled at the same time.<BR>Sorry, there are no old or current catalogues available, but we attach the <BR>list of parts for which we are still prepared to take orders. The items are made to order so delivery can be extensive. Prices are given <BR>on application.”

by hotrod on Sat Dec 01, 2007 8:50 am

The earliest I’ve seen is 1972 and the newest is 1976.

I’m not going to waste time on the dating research of this amp. It’s not a Marshall or Fender where most people get anal about that stuff. All that should matter is build/component quality, design and tone.

After the usual chassis removal, white spirit clean and Tolex nail brushing, I tested all the valves in the Ashton, with only 1 failure of an ECC83 pre amp heater blowing in the usual white flash.

The case and chassis cleaning makes such a big difference to the start of checking an amp to me practically, as you aren’t working in dust and cobwebs, and mentally – the personalisation factor – as it puts you in a frame of mind to want to get into the amp to discover what is going on electrically to get it working safely, once all the white spirit smell has disappeared after a day or so. I spray all the pots, jacks and switches with IPA at this point also so they are cleaned off and turn smoothly for bench testing. Brushing the circuit board with a soft brush and IPA also gives away any loose connections early on also.

It can reveal easily fixed minor issues like the heat grill needing re-stapling:

I’m going to look at the schematic with a bit of theory as I understand it, but that view may not be totally technically correct – you have been warned. This is helpful as a theory recap anyway, but this is a weird design that I haven’t seen before, so can only help to understand it before working on it.

This amp has a dedicated tone variable “Klipp” input channel – the lower jack in the schematic to triode2 pin 7 grid – which is why an extra ECC83 (valve 3) seems to have been added (a cathode follower? – output from triode 6, pin 8, 100k cathode to valve 4, triode 7 Long Tail Pair) to take the mixed channel 1 and 2 (tone varied) signals from valves 1 and 2 to valve 3. A 1M pot can be switched between clean signal or full distortion (the Klipp channel).

Looking at the valve wizard’s pages it seems valve 3 is a cathode follower of sorts, to retain tonal quality, bandwidth and frequency response from the Klipp stage:

http://www.valvewizard.co.uk/accf.html

“The AC-coupled cathode follower is normally used as a means of coupling a high output impedance to a low input impedance with minimal “insertion loss” (i.e., not much loss of signal amplitude). The cathode follower has very low input capacitance and a very high input impedance so it will not load down the previous stage, and very low output impedance so very little signal is lost even when driving a fairly low input impedance. The cathode follower is an excellent buffer stage for driving a tone stack or effects loop, or any circuit which would otherwise present a heavy load to a “normal” stage. In addition, the DC-Coupled cathode follower can also be used to produce a unique compressive quality, and is to be found in most of the classic amp designs. The AC-coupled version is not so useful for this, since the input couling cap prevents the flow of quiescent grid current. The AC-coupled version is therefore used as a tonally transparent stage, usually. The output signal is taken off the cathode and is in-phase with, or ‘follows’ the input; hence the name. The signal could also be taken off the junction of Rb, Rl and Rg, it does really matter which.
The load resistor is placed in the cathode circuit and results in 100% negative internal feedback taking place, giving very low non-linear distortion and a very linear response from very low to very high frequencies, making it especially suited to hifi applications. The only significant drawback is that the high level of internal feedback also limits the maximum voltage gain to slightly less than unity.”

This would be right as the tone stack follows this valve3 immediately after, as described.

A problem with this amp is the control panel markings are all worn off so I can’t tell which knobs do what without physically checking the wiring, as I can’t find a clear front photo for this Klipp model on the web either, but at least the two channels – normal and Klipp – are clear here:

I’m guessing the switch in the Normal channel is the one in the red path below, and the other two are in the Klipp, blue signal path I’ve drawn below. I’m guessing bass, middle and treble for knobs 2, 3 and 4 from left to right. The far right switch on my amp is marked “top” so a bright switch. The Normal channel switch will be the same.

From forums, it seems this amp can have a very high HT up to 700V! This can supposedly cause a problem with the screen grid on some new makes of EL34 valve that can’t handle that, maybe requiring diode chains to be added, but I will research this further once my preliminary circuit checks have been done. I have the Variac to use to check this amp also, so can bring it up gently, once I have replaced the 15uF cap that came taped to the chassis, which I have found the point where it has broken from – between the other remaining pre amp wiring:

If the prior owner had powered it up without the broken 15mF cap it would not have been a big problem, as this is a 15uF D+ filter cap only. The only thing I know wrong at present, except the failed valve, is the 1A mains fuse is blown. I’ll have to assume a reason for that and try to find it before power up.

This original cap (which I hope will still work) relates to the non original silver cap shown in Chambers repair here:


Turned out the cap seems ok but I can’t solder a wire to it – it won’t bond…I used a 32uF removed erroneously from the Mercury last year…

For interest and a recap, I will try to get an idea what each pre amp stage’s characteristics are, using the values in Chambonino’s schematic, by drawing the load lines.

Stage1 pre amp

The first ECC83 uses common values of 100k and 1.5k resistors for load and cathode:

This gives Blencowe’s basic example at 300V exactly so load lines of:

This gives a load and cathode current of 300V/100k = 3mA. This is the Y axis point, along with the max HT value of 300V that join to form the Load Line for the 100k load.

The cathode load line, gain and bias voltage (1.5V) is therefore:

This gives a pretty symmetrical swing for input signal voltages up to about 2Vpp for both triode 1 and 2 which are separate guitar inputs, each with a gain of about (220V-160V=60V)/1V = 60. This gives an output up to 6Vpp at the anode load for a 100mVpp guitar signal. Realise from the prior Post on Load Lines and the graph below that for an undistorted signal to pass to a second stage it is pointless having a symmetrical signal larger than 4Vpp input to an ECC83, even biased at -2V to allow a grid swing from 0V to -4V, as no more amplification can occur above this signal size at a triode input due to grid current limiting on one side (0V) and hitting maximum HT voltage on the other (-4V). Only distortion can occur after this, also because the grid curve distances are spread out unequally.

This is why FX boost pedals like Tube Screamers can work nicely at the first triode to boost the (100mV) signal from the guitar lead to push the triode 1 stage only, as there is plenty of headroom here still – up to 4Vpp max depending on the bias point and available HT. This may add more 2nd order harmonics at this point so the signal becomes richer in these, as the signal becomes more and more slightly asymmetrical over the grid bias lines. This may cause more distortion at a later triode stage if the FX pedal gain level is set too high for the original circuit design. It seems a balance over clean amplification and desired tone, moving through crunch to full distortion at earlier stages or not.

The key thing to understand is whether distortion is caused by overdriving the grid (too large an input) and/or asymmetrical bias point (warm toward 0Vor colder toward -4V biasing), or correctly designed clipping distortion using combinations of HT voltages, bias point and load resistor values – NOT from a larger input signal over grid limits.

A main thing to consider for each is to be mindful of why biasing is warm or cold – the closer to a warm 0V, then the easier it is to turn on the valve with a smaller input and have it draw current more easily. The more current flows the hotter the valve plate will be obviously, and the shorter its life may be due to plate/cathode/temperature stress. At 0V grid curve bias, the higher the quiescent current draw so more current is wasted without a signal. At a -4V grid, the triode is completely switched off even at 300V HT, so colder. This would require a larger input to switch the triode on, but the amplified signal would be completely asymmetrical seen below.

If a (pre amp) valve plate is visibly glowing red with no signal then there is something wrong with the HT or load resistor and it should be repaired or re-designed. The maximum AC coupled gk bias voltage is sometimes seen as 0V when a cathode is attached direct to ground with no cathode resistor.

Below – asymmetry due to bias point:

Stage 2 pre amp

For the second valve – triodes 3 and 4 that will have an amplified signal (about 6V max) input from triodes 1 and 2, there is a different gain design for each triode:

This seems a form of asymmetrical Long Tail Pair, with only 80V at the anode of triode3 because of the 22k and 47k voltage divider splitting the 300V HT between them, but not quite proportionally because each triode is asymmetrical in parallel also. Triode 3 has a 47k load and triode 4 has a 47k+10k = 57k load.

If you wanted to work it out roughly, use the spec sheet value for the EEC83 figures for each anode plate resistance:

Typical

characteristic:

Ua = 250 V

Ug = -2 V

Ia = 1,2 mA

S = 1,6 mA/V

Ri = 62,5 k?

??= 100

This would give a triode 3 resistance chain of 62.5k + 1.5k + 10k = 87.5k

This would give a triode 4 resistance chain of 10k + 62.5k + 1.5k + 10k = 97.5k

The total valve resistance is roughly 1/87.5k + 1/ 97.5k = 1/R = 1/0.00001577. R = 46k.

This value is in parallel with the 22k, so these give a resistance of about 1/22k + 1/46k = 15k

Simplified, this gives:

This is in series with the 47k triode3 load, so the proportional voltage for this divider chain is 46k/15k = 3.1:1 or 300V/ 3.1 = 95V or so. The measured schematic value is 80V at the anode as written.

Close enough to see how that divider works.

Ok, the other major design difference here I haven’t seen before is that this stage only amplifies the output for triode2 (blue). Triode 1 bypasses it completely (red) and goes straight to valve 3, triode 5.

Triode 4 grid is fixed to 0V ground via the 100nF cap to pin 7. This is the fixed reference for the LTP at this valve (orange line). Remember, an LTP amplifies the difference between both grids.

The gains and clipped stage plots for this valve2 may be something like those below once the calculations are done for each triode load resistor.

The schematic value is 80V so maybe I’ll just work out the quiescent values for this triode3.

The resistance of the triodes 3 and 4 in parallel similarly worked out as above.

Plate resistances and 10k load in parallel = 1/62.5k + (1/72.5k) = 33564 ohm.

This is in series with the 11.5k cathode resistance = 45k ohm

Total quiescent valve current = 80V/45k = 1.7mA

11.5k cathode resistance x 1.7mA = 20.4V cathode voltage.

20.4V/6.6 ratio = 3.1V bias across the 1k5



Well even if the bias point of -3.1V is wrong, it seems if you want distortion, you will get it here no matter what the bias is, as there would be no cathode current at all for any grid signal above 1.5Vpp, and there may be up to 6Vpp from the first pre amp because of the up to 60 gain from valve1.

A change of 1V grid voltage gives a change of anode load voltage of about 70-35V= 35V so a gain of about 35V/1V = 35. This signal gain at triode3 is superimposed at the load resistor of triode4 so a load line for that triode could be drawn also, but it is still at around 1.7mA but with a slightly lower HT due to the drop across the 10k load, so little point. You get the general idea hopefully.

Now it seems you can have the output from triode 2 AND/OR triode 4 passed on AND mixed with the output of triode 1 at the grid of triode 5 depending on the position of the switch KlippSW. I think this is the “Klipp” switch that gives the full distortion tone, but will have to check the wiring to mark the panel. Both red and blue paths seem to have optional 100pF bright caps switches also, as there are 5 switches on my unmarked panel. 1 mains, 1 standby, Klipp switch and so I guess, these 2 tone switches – 1 in the Normal section and 1 in the Klipp section. This implies that one knob in the Klipp section may be a 1M pot “mix” control between clean (red) and distorted Klipp (blue) signals, and the other a volume/mix pot to triode5 for that blue mix with the red clean channel.

The 3rd pot is another volume/mix pot for the clean red channel and triode5 volume which must be pot 5 on the panel and pot 1 is the 5k NFB presence pot at valve 4 LTP cathode.

Stage 3 triodes 5 and 6 and tone stack

While I’m looking at the tone controls, I can try to identify each using maths from the filters Post.

A large cap passes lower frequencies, so the smaller 250pF cap with the 250k pot must be the treble. The cut off frequency is:

F = ½ Pi RC = 1 / 2 x 3.142 x 250 000 x 0.000 000 000 25 = 2545 Hz Treble. All frequencies above this pass.

The next is the 22nF with the 1M pot:

F = ½ Pi RC = 1 / 2 x 3.142 x 1000 000 x 0.000 000 022 = 7.2 Hz Bass. All frequencies above this pass on when full up. It’s the Bass pot. Wow, makes a bassy amp maybe!

The mid cut off is:

F = ½ Pi RC = 1 / 2 x 3.142 x 25 000 x 0.000 000 022 = 289.3 Hz Mid. All frequencies above this pass on when full up.

The Presence pot cut off is:

F = ½ Pi RC = 1 / 2 x 3.142 x 5000 x 0.000 000 1 = 318.3 Hz Presence. All frequencies above this are NOT negatively fed back (as they pass to ground when full up) so NOT suppressed at the grid of triode 5 (pin2) by working in opposition to the signal at triode6, pin7, so are relatively amplified to others below this frequency.

This Klipp pre amp design is quite an interesting and maybe unique tone arrangement for an amp. I wonder if Laney patented it?

Stages 2 and 3

A thing I think happens is all 3 pots will affect not just the relative signal mix, but the signal level at the grid of triode 5, so the gain there will be a combination of all three pots (if a 2 guitars are plugged in – one at each channel). Also, if both red and blue signals here are about the same level, the maximum level at the grid will depend on whether both red and blue are in phase or not. I will check this when I bench the amp.

As all triodes with an anode load are inverters, my initial guess on phases is below:

As signals in the same phase add together then the inputs at triode5 grid will ADD if I’m right. Also note that valves 2 and 4 are effectively Push-Pull LTP stages as these cathode tails are shared between both triodes of each valve. Note the anode of valve3 triode5 is connected to the grid of triode6. This is in line with the unity gain of this impedance matching cathode follower, so it’s only a small output from the anode being fed to the grid with no DC decoupling capacitor.

This valve 3 is the mixer for both Klipp and Normal channels.

So what have I got so far for the Stages 1, 2 and 3 summary?

First, two clean signals from both valve 1 triodes.

The red path passes unmolested to valve3, and the blue path is split at valve2 to form one clean and one distorted signal that can then be re-mixed together.

This distorted and/or clean signal mix can now be further mixed with the red clean signal from triode2 valve 1 output or have two totally clean signals added together (via the Klipp distortion switch off) at the grid of valve3 . Either clean or distorted blue signals can be bright switched also, as well as the red clean signal bright switched. The tone is described by some as fuzzy distortion:

by klipp on Thu Nov 29, 2007 9:40 pm

the KLIPP “fuzz” feature is actually a Hyper-Compression kind of circuit.
It Does Fuzz, but its a Wooly kind of fuzz / Saturation, the louder the volume, the Bassier / wooly-er it gets.”

Sounds horrible so far! Haha. We’ll see – its valve “fuzz” not square wave transistor noise.

I don’t know what he means by “hyper-compression” – all distortion is a form of compression/limiting. A signal gets clipped because it can’t rise in amplitude over a certain point – its voltage ceiling – as it saturates passing maximum current, so there is no headroom left by the limiting factors of max HT to rise to, or the bias current cut off values at the grid.

Tonal Options Summary

1: Three signals mixing/volume pots (2 x clean, 1 x fuzz).

2: Three x bright switches (1 per signal).

3: Four tone pots for the combo red and blue signals – bass, mid, treble and presence pots.

Stage 4 LTP pre amp

This stage includes the NFB presence control mentioned above.

I think it’s safe to assume that after all the effort put in to tone creation in stage 2, 3 and the tone stack, that the stage 4 LTP will just attempt to faithfully reproduce those tones without much further colouring except for the presence knob, and provide the correct output impedance and signal levels to the power stage.

When checking the circuit I found a very different value of NFB resistor than the schematic 100k linking to the 16 ohm OT secondary – a 4k7 marked, but 6k2 value resistor*:
*I was being a dick here – this was not a very different value – I was measuring the OT 16ohm secondary in parallel with it – DUH!

Also, I found the slave line missed off the schematic. I checked all the connections manually that are marked red. I have nearly finished the whole amp now, so close to mains testing in the next day.

The main concern is always whether the PT, OT or choke appear to work, as they are the most expensive to replace, and looking at the cost of Partridge Transformers spec info to find equivalents, it’s a potential nightmare to replace them. All seems ok for now with all the OT windings showing nicely on the scope for each ohm switch settings – 4, 8 and 16 ohm:

I have tested the amp to a thorough level that I’m content that there are no major faults. I marked the diagram in red as I went through each section:

I have re-wired the really bad Earth/ground/fuse wiring – I should have taken a picture before correcting it, but it was wired as the schematic with the PT sec fuse connected to the rectifier negative:

It will be wired as below with the fuse moved as per Blencowe’s fuses text:

“Before the rectifier, in series with the power transformer HT secondary (recommended): This is the best place for the HT fuse as it protects everything, and is also in the AC part of the circuit where most fuses are designed to work. 
With hand-wired amps this is perhaps less practical with a two-phase rectifier since it requires two fuses, one for each leg of the transformer secondary.”

http://valvewizard.co.uk/fuses.html

The mains earth has a dedicated connection bolt now that I drilled a hole specifically. There are NO chassis ground points on this amp I can find – not even the heater wire centre-tap as per the schematic?

The circuit ground connected directly to mains earth at the fuse holder! The consequence of this is if the fuse blows then the DC side shuts down ok as the DC circuit is broken, but there is still PT secondary AC at the rectifier diodes.

I need to test and replace the neon bulbs, rewire the PT sec fuse holder, source two new 1 and 2A fuses for the PT then I am ready to start the first Variac mains test. If the two Carlsbro amps have the correct fuse values I can use those as they are the older 32mm x 6mm types also. I bought some on Ebay as my useless local store doesn’t stock them.

This Post is getting large so I will do another for the mains testing.

Leave a Reply

Your email address will not be published. Required fields are marked *