Category Archives: Repairs

AX84 – P1 Extreme – grid stoppers?

The problem

While playing my P1 Extreme over the last weekend I noticed an odd sound when letting a note sustain. Even when run clean, the amp would produce a fuzzy/distorted version of the note and then within a second or so the whole amp would cut out.

Investigation

It didn’t seem to matter which guitar I used, or which note I played. This has meant the problem has been hard to reproduce. Still trying though.

I asked members of the Hoffman forum and they thought it was caused by oscillation, and suggested the following:

  • “Chop-stick” the wires inside the amp. i.e. use a non-conducting device to reposition wires in relation to each other. The proximity of low level signal wires (guitar input) to high voltage wires (power supply) can lead to problems.
  • Make sure the grid-stoppers are located as close to the tube socket pins as possible.

Upon reading the forum’s suggestions I realised that I knew nothing of what grid-stoppers were meant for. Also, looking at the amp’s schematic I saw that the input stage tube, V1, doesn’t even have a grid-stopper. Here’s an excellent discussion on why this might be, in the case of an AX84 amp. Secondly, the power tube’s grid stopper is located on the main board, with a wire linking it to the tube socket. This would appear to be an opportunity for improvement.

Here’s what I’m going to do:

  1. Digest this article about grid-stoppers on the Aiken Amps site.
  2. Get myself a chopstick and carefully poke around the amp while trying to reproduce the problem. “Lead-dress” might be the issue, and is relatively simple to fix.
  3. Move the existing power tube grid-stopper to be soldered directly to the tube socket.
  4. Add a grid-stopper (10k – 65k value resistor) to the V1 grid.

Work (not) done

I read the article about grid-stoppers and also some more specific AX84 project stuff. These articles were discussing why the P1eX doesn’t have grid stoppers on the pre-amp stages, and why the grid stopper for the output stage is on the main board, and not soldered directly to the pin on the tube socket. One poster reasoned that the pre-amp stage grid stopper was “missing” due to the fact that the original designer didn’t have RF problems in his house. THere is a note about this on the schematic. Who knows?

As for the output stage? I didn’t move the resistor. I didn’t make any modification to the output stage at all as I could not reproduce the problem. What I think I saw/heard was oscillation caused by resonance from the speaker cab. It was literally vibrating itself into oscillation.

However I did replace some of the cheap/crap/oversized Monster shielded cable that I had originally used. Now it has some decent shielded cable bought from Hoffman Amps. It’s listed here asMini shielded cable RG174″ I got about 5 feet, so that’ll last me a while. This also meant I was able to rewire the input jack properly and take advantage of the Cliff Jack’s switching feature. When a guitar cable is unplugged, the tip is now shorted to ground. It makes for a silent amp when nothing is plugged in.

I also added a shield for the pre-amp tube.

Outcome

So my P1eX sounds great again. My son and I had time for a jam yesterday, so I was able to use it for a couple of hours without any issues at all. These days I use a a 6L6 for the output tube, that I bought from Hoffman at the same time as the wire/shields. It doesn’t sound all that much different to the EL34, but I haven’t trully cranked it which is where I may hear differences.

Also I noticed the power transformer (PT) was warmer than I had noticed with a 6V6 output tube. After some brief reading I decided to not worry about it at all as it’s not so hot that I cannot keep my hand firmly on it with no problems.

Update!

I recently made a few little changes:

  • Switched back to a 6V6, and changed the output tube cathode resistor to a single 680 Ohm.
  • Replaced the last of the crappy Monster cable.
  • Moved the grid-resistor to the tube pin.

This led to the amp not producing any sound at all, which was worrying. I decided to check all the voltages, record them and ask a question in the Hoffman forum. You’ll find the thread, here: Odd voltages in P1 eXtreme.

The outcome was that I found a dry solder joint right where the cathode resistor/bypass-cap attach to ground. Once I removed old solder, and re-did these turrets all the voltages returned to within spec.

Dry joint in cathode ground

Dry joint in cathode ground. I removed and re-did thes joints inside the orange box.

Here’s a shot of the grid resistor in its new location right on the grid pin of the output tube:

 

Grid resistor wiring

Grid resistor wiring and heatshrink

Eye-candy

I noticed some blue glow in the output tube. I’ve read enough to know not to worry about this either as it’s apparently normal. But I used it as an excuse to take a long exposure shot of it in relative darkness. This was a 20 second exposure…

An Electro Harmonix 6L6 in an AX84 P1 eX

An Electro Harmonix 6L6 in an AX84 P1 eX....ain't she purdy?

Advertisements
Tagged , , , , , , , , , , , , , , , ,

Bias circuit capacitors

Bias circuit caps

Here's the new capacitors from the bias circuit. You can also see the 'pegged' bias pot. I'll be replacing that too.

Next on my list of ‘replace old components’ are the electrolytic capacitors present in the bias circuit. Doing this was an ENORMOUS pain in the arse as the PCB in the 2204 is by its very nature, soldered underneath the board. This meant I had to un-bolt the board from its 6 posts (luckily I had a nice screwdriver/wrench for this) and attempt to lift it away from the chassis. Clearly this had never ever been done before and I’m sure the board was first bolted in and then wired up to the tubes/pots/power. Marshall did not have the foresight to make the wires long enough to be able to lift the board and work on the underside. I made a note (just with a piece of paper) of the wires emanating from the edge of the board, and where they were soldered to. This meant I could de-solder a few of them, thus providing better access. It only took a few lengths of wire before I could turn the board over enough to get access to the bottom.

Once done, I hooked everything back up and fired her up. All was well, so I went ahead and checked the bias, which hadn’t changed at all. You can see in this photo the bias adjustment pot at the end of it’s track, so to speak. I will be checking the value of this in the future, and perhaps replacing it.

Cap job

Discharging the old Davy capacitors

Discharging the old Davy capacitors with my new tool. Note the horrible quality soldering, courtesy of Marshall.

This past weekend I was able to replace the larger 50-50 caps in my amp. It meant I got to use my capacitor discharge tool which I’d been toying with for a while. Soldered between two alligator clips, I have a 5W 5k Ohm wire-wound resistor. I left them clipped on each positive cap terminal for a few minutes. That seemed to do the trick; my test was to do “the screwdriver trick” and short them to ground. There was no “loud pop and sparks” so my tool was successful.

Original Davy 50-50 cap being removed

Original Davy 50-50 cap being removed

So I de-soldered the old wiring and removed the old clamps. It was easy to put the new ones in and re-attach the wiring. I had taken pictures along the way,so it was trivial to get the wiring back in the right place. History has shown that my memory is close to useless, so I didn’t even attempt to hold it all in my head.

You can see the underside of the main PCB through the hole in the chassis (see left). I left this cap, the one sitting next to the choke,  sitting quite high in its clamp (see the picture at the end of this post) as I was paranoid about shorting the wires against the PCB.

Two of the new caps in place

Two of the new caps in place.

You can also see in the above pictures that Marshall wasn’t too picky about employing people that could solder particularly well. Their preparation seemed to be the key…i.e. when they stripped insulation off wires they did not bother to spend half a second to twist the strands together before tinning it. I had to replace a few wires as the original wire was not going to hold up too well.

Once completed, I fired her up and checked the bias. All was normal, and it appeared to have greatly reduced the amount of hum…especially when cranked. Success!

Next step will be replacing the electrolytic caps in the grid bias circuit.

New filter caps installed

New filter caps installed

Tagged ,

Bias adjustment prior to replacing capacitors

This last weekend I was able to begin the process of restoring the innards of my 2204. The very first thing I did (after discharging the filter caps properly) was solder a 1 Ohm resistor between the cathode and ground of the EL34 output tubes. Once complete I was able to check the bias of the existing setup by measuring voltage across that resistor. Ohms law provides that he voltage I measured was equivalent to amps. It was way down at 19 mA.

So I referred to this Eurotubes biasing page which told me:

The formula for calculating the bias is the dissipation of the tube divided by the plate voltage which gives you 100% and then you can bias anywhere between 65% up to 85% of this.

The bias components on the 2204

The bias components on the 2204 board. They’ll need replacing soon.

My plate voltage is 400 so using 25 watts as a basis, 25 / 400 = 0.062. 70% of that is 0.062 x 0.7 = 0.043 = 43 mA. Tweaking the bias trimmer gave me 39mA on one tube and 41mA on the other. A reasonable match.

This didn’t seem to affect the tone using the original tubes and filter caps. I recorded the resulting ‘tone’ at a known setting that gives me the Marshall sweet-spot tone.

Two things I noticed that will need investigation:

  • The bias adjustment trimmer was almost at the end of its travel to give me the value I wanted. This may mean the trimmer is close to death?
  • Why was it set at 19mA in the first place? Is this also an indication that my plan to replace the bias components really must be done soon?

UPDATE from the future! March 2015

I replaced the output tubes recently and biased it the same way as above. When playing the amp in anger at a rehearsal, I found that the amp breaks up into quite pleasing distortion much earlier than I expected. It is indeed a great sound, but I’ve lost the strident clean tone I would normally get. There’s one song we do that needs that…for my Telecaster and a tremelo effect. I now suspect that the new tubes were simply born this way, and that I need to change the bias so the amp is running a little colder.

Look at this post: http://www.mylespaul.com/forums/squawk-box/31210-getting-jmp-2204-bias-within-range.html

FWIW my rule of thumb for JMPs is 28-32 mA for EL34s, 34-36 mA for 6550s. I sometimes go a little hotter, but to my ears there’s generally no sound improvement for the shorter tube life. I generally shoot for something like 15 watt plate dissipation.

Most tube dealers recommend you bias really hot. I wonder why?

This implies I biased it REALLY hot at 41mA. We are going into the studio tomorrow, so I’m going to rebias first thing, and see if I get my ultra clear ‘barking’ clean tone back.

UPDATE again: Dec 2015

It is worth stating that I was indeed able to get my loud and clear tone back. My amp was indeed biased way too hot. I can still get good 2004 grinding tones by using lots of pre-amp (I might put it on halfway) and dial the master volume to suit the loudness I need. But mostly I run mine with the master all the way, and adjust the pre-amp knob to suit loudness. There’s a sweetspot where I get a great strident clean tone that takes pedals very well.

Tagged , , , ,

More components to replace

After some discussion on the Hoffman/EL34 forum, it seems that as I am going to replace the filter caps in my Marshall 2204, it’d be a good idea to replace the “bias components”. This basically means that I should replace all components between the power supply and pin 5 (grid) of each EL34. This little circuit includes a couple of electrolytic capacitors, so the same rules apply…if they’re old, replace them. I’m going to replace the resistors in the circuit with metal film resistors. I don’t see any harm in doing so, and it’ll be good practice.

Master Model 50w Mk 2. Lead - my amp

The schematic I’m using appears accurate when compared to photos of my board. That in itself is giving me practice in identifying the right components, and tracing the circuit visually. There’s a cognitive leap from schematic to physical layout that takes getting used to…like translating a UML diagram to actual source code.

You’ll see a section above the power transformer that is sort of labelled ‘BIAS’. Note the labels by some of the resistors, RG, RB1 and RB2. These seem to be variants for the European market. This initially confused me, as the parts on the actual board didn’t map to the schematic. That was until I remembered I’m British so my board has ‘all’ in ‘all but USA’ component values (see the little box above V4/V5 in the heater wiring section?). I was seeing 150k on the schematic, but finding 200k on the board.

Luckily I made this discovery before I ordered all the remaining parts from Hoffman.

Tagged , , , , ,

Replacing/discharging capacitors?

While searching for stuff about my Marshall 2204, I happened upon a fantastic source of information; The Hoffman Amplifier forum, hosted by EL34 World. I saw that there was an ‘introduce yourself section’ so I posted this: Hello from Chicagoland

A forum member was kind enough to point me in the right direction (the Hoffman Amp store) after suggesting I replace the ‘electrolytic capacitors’ rather than replace the output tubes.

So from Hoffman, I’ve ordered 3 50-50 500v ‘can’ capacitors. The logic being, the Daly capacitors on the amp are original and therefore very old. They might be the source of all the noise/hum when the amp is even slightly cranked (in the sweet spot, so to speak). I’ve also ordered parts to make a capacitor discharge probe. It seems there’s no consensus on how to do this. Techniques vary from taking a screw driver and shorting the caps to the chassis. This can’t be particularly safe, although it’d be dramatic. Other methods involve essentially the same thing, but using a resistor in series to slow the rate of discharge. But nobody can agree on the value of such a resistor. I have found these variants:

  • 100k Ohm, 5W
  • 470 Ohm – 2k Ohm, 2 Watt

I have ordered a 5k, 5W wirewound resistor. And I already have some clips and heat shrink with which to make the tool.

I might eventually make one of these though: Capacitor Discharge indicator. It implies that “finishing it off with the ol’ screwdriver” is still a good idea once voltage drops below lethal levels.

UPDATE!

My good friend Tom Reid has some insights on this:

My standard discharge tool is a very large Craftsman screwdriver. It worked
well for picture tubes over the years. The added resistor to a discharge probe creates a current through to discharge procedure. For best results, discharge through a standard incandescent light bulb, as the filament is both resistive, and inductive so as to absorb the shock.

UPDATE AGAIN!!

Finally I found a page where someone quantifies the value of the resistor one might use:

It’s smart, then, to discharge the unit deliberately, through a resistor equal to about one ohm per volt of charge.

Tagged , , , , ,

EL34 tube prices

After my previous post outlining how I found that I had a bad output tube, I naturally needed to replace them. My 50W 2004 has two EL34 tubes in the output section, so I must replace them as a matched pair. As I haven’t bought tubes since ’94 I know nothing about the best places to get them. Here’s what I found online:

  • Tube Depot – Wow! Their prices range from less than $30 for a pair of Electro Harmonix to more than $450 for a single Mullard tube. That’s just ridiculous.
  • Tubes and More – Much more reasonable, and they have Groove Tubes too.
  • The Tube Store – Not many in stock, but good prices
  • Sweetwater – All kinds of tubes, with ‘larger company’ return policies.
  • Audio Tubes – More insane prices.

It looks like the market for “vintage tubes” suffers from the same bizarre and irrational pricing justifications that I’ve only ever seen in vintage guitars and so-called high end audio components. There are plenty of snake-oil salesman trying to get us to pay stupid prices for a green marker pen to colour in the sides of your CDs. Or the type that’ll argue that analogue technology as a recording or playback medium is objectively better than digital recording. Here’s a great rant about those types.

In this case I’m going to go with what Mitchell recommends in this book, and get a pair of Groove Tube EL-34M from Sweetwater. They’re supposed to be a “reproduction” of “the classic” Mullard XF-2 “design”, which is apparently “much sought after” (here’s the press release). Let’s face it, my 2204 sounds incredible with the so-called crappy Russian tubes it’s had since ’94, but what do I know.

Tagged ,

My 2204 – Initial Exploration

This post talks about how I did an initial exploration of my Marshall JMP 50 Watt ‘Master Volume’ (model 2204). As a continuation of my previous post, here’s an update:

  • The multimeter arrived! It’s incredibly basic (it’s even a manual ranging device, which means I must be careful) but very well built.
  • I couldn’t get hold of a 1k Ohm wire-wound resistor to make my own capacitor discharge doohicky, but unless I was going to be desoldering components I left that for later.

So my goal was to use the voltage chart contained on page 178 of the Tom Mitchell book to check my own amp. I was hoping that my amps values would be “within 10%” of the published values. To do this I needed to measure voltages “at the pins” of the various tubes in the amplifier. That meant I needed access to those pins in a safe manner. The last thing I wanted to do was injure/kill myself so I followed accepted procedure by making sure I only ever used one hand to place the red probe onto the pins. The black probe was clipped onto the chassis (so I didn’t even have to use my other hand to hold the black probe, which was behind my back, and therefore unlikely to touch anything).

2204 and Multimeter

The 2204 chassis on it's back ready for probing.

I unscrewed the back panel of the amplifier, and then removed the 4 screws that secured the chassis to the cabinet. When the last screw came out, the chassis kind of plopped forward due to the sheer weight of the transformers. Probably 90% of the weight of this thing is those transformers. I then carefully picked up the chassis from inside the cabinet and placed it upside down on top of my 2×12 cabinet (which was itself on it’s side). This gave me easy access to both sides of the chassis, and non of the tubes were touching the supporting surface, and therefore not being strained mechanically. It occurred to me that I really should build a support frame if I was going to do this on a regular basis; something that would allow me to flip the chassis about easily without having to risk touching components inside.

The rabbit's nest of wiring

The inside of my 2204 wasn't exactly pretty

There was more risk involved as the inside of this particular amplifier wasn’t exactly pretty. After looking at the insides of other people’s custom builds, my 2204 looked like it was thrown together correctly but with little regard for “craftsmanship” or “functional beauty”. I was hoping for this kind of attention to detail, but was disappointed to find a rabbit’s nest of wire and seemingly shoddy soldering. Firstly I made sure the black probe was safely clipped to the chassis, and that anything and everything was out of harms way (including some of my pets which were messing around on the floor. I made sure they were out of the room). Secondly I attached the speakers, as it’d be very bad to power up the amp without them; the speakers are “part of the circuit”. I then turned on my multimeter and dialed it to the highest DC voltage setting.

2204 and the multimeter

You can see the black probe clipped to the chassis, on the left. The multimeter set to it's highest DC voltage setting. The Mitchell book and the 2204 chart ready on the right.

The assumption here is that all the voltages measured are “in relation to ground”. It’s not like testing individual components where you might measure the voltage drop across a resistor, or test for continuity. It seems all voltages appearing on schematics or charts like the ones in the Mitchell book, are all about their relationship to ground, hence clipping the black probe to the chassis and simply touching the red probe to the appropriate pins.

So, now I was ready to measure something! I thought I’d start with the amplifier on standby as I knew from reading that all this did was provide power to the pilot light (in this case, inside the rocker switch) and power to the heaters (filament) inside each tube. I could see that each tube was indeed glowing and the pilot light was on, so with my red probe I headed for pin 9 on the first ECC83 with excitement. Nothing. Nothing at all. Zero. I tried pins 4and 5 , which turns out are soldered together. Nothing on those either. So, first lesson of the day….the book goes on to describe the voltages expected on the heaters of a 12AX7ECC83…3 volts. Ok, so where’s my 3 volts? Well, Mitchell is very clear to point out that the voltage on heaters is AC, not DC. That requires a different multimeter setting. Once set correctly, I got a nice steady 3 V AC on all my tube’s heater pins.

Right, so back to the other pins.This was where it got weird. I was getting great numbers as I made my way around the ECC83s. It all made sense until I got to the EL34 tubes. According to the Mitchell chart I was to expect a voltage on pin 6. But it was clear from looking at the pins that nothing had EVER been soldered to those pins. I did some poking around and found that pin 6 on an EL34 is officially “no connection”. Mitchell’s own diagram in the book agrees with this page.

I was able to measure DC voltages in my 2204 that were within 10% of the ones in the book, which I took to be successful. I also took this to mean that nothing was fundamentally wrong with the way voltages were in the amplifier, so my odd metallic resonance may be coming from something mechanical. I decided to take a break so I turned off the amplifier and unplugged it to have a think and perhaps read some more.

2204 in "tube tapping" mode

With the amp in more normal repose,with a guitar attached to reproduce the metallic resonance.

One particular test in the Mitchell book describes the act of tapping the tubes with an insulating object (such as his much touted Sharpie). So after getting the amp the right way up and turning it back on, I plugged in a guitar and fiddled with it to get it loud enough to reproduce the errant noise. I then found a nylon pen of suitable length and proceeded to gently tap the output tubes. Sure enough, one of the EL34 tubes made a very distinctive “thonk” when I tapped it. The other was completely silent. I believe I have found the problem and I think it simply comes under the term “bad tube”. My plan today is to buy a matched pair of EL34 tubes. Mitchell unconditionally recommends Groove Tubes or Mesa Boogie tubes. I’m going to shop around for a decent price obviously, but will stick to these recommendations for this first effort. That’ll be another post as this one’s way too long already.

Tagged , , , , , , , ,

First steps toward repair

As I began poking around looking for pointers on how to even start repairing my Marshall, I found a plethora of web sites that have already proved useful. They are in the very least interesting from my beginner’s perspective:

  • The AX84 Project – They have a forum (or as they refer to to it as a BBS…how quaint!) full of fascinating discussions on pretty much any topic. The level of quality here is wonderful. Very high-signal to noise ratio. Secondly there are tried and tested designs for low powered 9well, and high powered) tube amps. I will be buying one of these kits soon, but I must be disciplined and repair by Marshall first. I know that this experience will inform the next.
  • Tom Mitchell’s Book – I bought this the other day, and have already devoured it. Not that I claim to understand it all, not even slightly, but there are nuggets of genius scattered throughout the chapters.
  • EHN’s AX84 page – Great coverage of what can be done using the AX84 project’s designs as a springboard. There are some great photos to go along with the sound samples.

So, the last few days I have been reading the Mitchell book in preparation for the 2204 repair. Also, to that end:

  • Ordered a digital multimeter. I went for the B&K 2703C, based on Mitchell’s recommendation.
  • Went to Radio Shack and bought 60/40 solder, some heat-shrink, some alligator clips and some 18 gauge hook-up wire. With this I will build a simple tool for draining the power supply capacitors.