The "Ultimate Attenuator"

ICBM

I've just acquired one of these, thanks to @thisisguitar who had a major problem with it - it blew up his amp :-O. Having known a bit about the "Ultimate Attenuator" and fixed a non-working one previously, I was not surprised. Luckily because it was a well-made amp, it only fried a valve and not anything more serious…

Firstly, the "Ultimate Attenuator" is not an attenuator - it's a re-amp system in a box. Basically this means that instead of reducing the amount of power the amp sends to the speaker by varying the proportion that goes to either the speaker or a dummy load, it absorbs it *all* in a dummy load and then re-amplifies a controlled amount of it via a solid-state power amp - hence why it needs a mains supply - but the "Ultimate Solid-State Re-Amp System" perhaps doesn't sound so good from a marketing point of view. The maker also claims that the re-amplifier is not simply a solid-state power amp but a "proprietary circuit".

The first bad thing: although the "Ultimate Attenuator" can supposedly be used with any amp, the dummy load is 30 ohms. For anyone remotely familiar with valve amp impedances (typically 4, 8 or 16 ohms) and the need to match them to the speaker/load, this should sound like an odd - and bad - idea. It is. Even though a real speaker has a rising impedance with frequency so that its average impedance is somewhat higher than the nominal value, a mismatch of more than double is still too high. In other words 30 ohms is safe with a 16-ohm amp, but pushing it even for an 8-ohm one (which is how thisisguitar was running it) and definitely not safe with anything below that. The usual result of an over-high impedance is arcing in the amp, often in the valves but sometimes at the valve socket or if you're really unlucky, in the output transformer - which will destroy it. It can also cause valve screen grid failure, which will also kill the valve. If anything valve amps are more tolerant of a too-*low* load than too high.

I did test it as it was - it works, although the claims for being the most transparent attenuator on the market are nonsense. It sounds flat and buzzy compared to most standard attenuators I've tried, including the Marshall Powerbrake I have and compared it to. If anything, it sounds exactly as I would expect from an over-high impedance load… hmmm. So it's neither the safest *or* the most transparent attenuator on the market.

So given all this, I decided to find out exactly how it works and if it could be improved at all, or whether it was just going to be a box of spare parts. More to follow.

ICBM

It's not only badly designed, it's astonishingly badly made. The last one I worked on wasn't great, but this one is worse. So bad that it can't actually be considered safe either to the amp or the user, and would not pass any safety certification I know of anywhere, let alone in the EU. It's DIY prototype standard at best.

The very first thing I noticed was this missing screw in the bottom. (Top right)



Which does not correspond to a hole inside, there's a bit of metal in the way. This is why:







That lash-up is the dummy load. Because the holes in the casing don't correspond to the holes in its "frame", it's held in by one self-tapping screw which is not even tight, so the whole thing can rattle about and stress the wiring which is attached to it.

I took it out and it looks like this. Yes really.





Next I had a look at the mains wiring.



The cable clamp is crudely fitted through a rough hole hacked in the casing and then glued up, the heatshrinked joint on the live wire is actually slightly punctured at one point due to the solder lump underneath being spiky (hard to see in the pic) and the earth connection consists of a crudely bent bit of copper wire trapped none too tightly under one of the transformer bolts with the earth cable tacked to it. (Note there are no shakeproof washers anywhere in the whole thing - although there are bent copper wire loops which the builder presumably thinks do the same job - they don't.)

On top of that the fuseholder is an inline type - functional but not ideal - and although the unit is labelled as 240V on the outside, the extremely cheap-looking transformer is labelled as a dual 110V winding - ie a 220V one. It works, but again is not good practice.

None of this would meet any safety standards and some is potentially dangerous.


(Edit - old Photobucket pics replaced.)

ToneControl

can't you just trim the cone off a driver and use the voice coil as a load?

ICBM

(Continued from part 1)

From an amp point of view, it's also none too safe. Bear in mind that any risk of something going open circuit when there's a cranked amp connected to it, will result in a very high chance of major damage to the amp.

Firstly the amp-in and speaker-out jacks are very cheap flimsy far-east types which can be prone to coming apart and I would not really trust for a speaker connection. Secondly, the bypass switch works by running the signal though both switch contacts in series - the typical true-bypass arrangement, but which means that a bad contact in *either* will give an open circuit - there's a better way of doing this.



(I think the taped fan connections might actually be due to a previous repair, so I won't blame the builder for that.)

Thirdly, the dummy load circuitry is full of badly-soldered connections, flimsy soldering and floating bits of wiring which could easily break, again leaving the amp with no load.



I took the dummy load apart and found that almost every one of the joints hadn't "wetted" properly to either the resistor or the wire.





Just appalling workmanship.


(Edit - old Photobucket pics replaced.)

ICBM

ToneControl said:

can't you just trim the cone off a driver and use the voice coil as a load?

No, it will have almost no power handling capacity. The voice coil needs to be cooled by movement, without that it's just a coil of wire and will quickly burn out.

xscaramanga

Have you tried any of the other expensive attenuators from the USA that gear snobs rave about, ICBM? 

I've got a Motherload Elemental, which is great, and a Weber Minimass which I also like. I'm curious about things like the Alex's attenuator, the Ironman and various other posh ones on the market. To be honest, I struggle to imagine they could be a lot better than the Motherload because with the amps I've tried, it seems excellent at all volumes.

ICBM

jonnyscaramanga said:

Have you tried any of the other expensive attenuators from the USA that gear snobs rave about, ICBM? 

I've got a Motherload Elemental, which is great, and a Weber Minimass which I also like. I'm curious about things like the Alex's attenuator, the Ironman and various other posh ones on the market. To be honest, I struggle to imagine they could be a lot better than the Motherload because with the amps I've tried, it seems excellent at all volumes.

I also had the Motherload Elemental, which is the best I've come across so far - I only sold it because I didn't really need the recording/DI options and it seemed a lot of money to have tied up in it. I haven't tried the Alex or the Ironman, but a friend (the owner of the other Ultimate I worked on, now sold) owns the Alex and thinks it's great. I wasn't impressed by either the build or sound quality of the Minimass I tried, but it was the original angled-box one and not the newer red one, so hopefully they've improved.

dunstanmus

Awful piece of kit as you have shown. How do you like the Fryette Power Station?

Danny1969

So it looks like 4  x 30 watt resistors in parallel presenting roughly 8 ohms, that absorbs the amps output and then a small AC audio signal is tapped off and fed into a crude power amp ?

It really does look at awful lashup ........... those resistors need a substantial heatsink in order to operate at their 50 watt rating, not the  I.C rads they have used there

ToneControl

ICBM said:

ToneControl said:

can't you just trim the cone off a driver and use the voice coil as a load?

No, it will have almost no power handling capacity. The voice coil needs to be cooled by movement, without that it's just a coil of wire and will quickly burn out.

I thought some people ran them in a can of oil?

ICBM

Danny1969 said:

So it looks like 4  x 30 watt resistors in parallel presenting roughly 8 ohms

No, the resistors are in series-parallel to give 30 ohms. But you're right otherwise, which gave me an idea of course! (More later in the rebuild section .)

Danny1969 said:

... that absorbs the amps output and then a small AC audio signal is tapped off and fed into a crude power amp ?

Which brings me on to the next bit - which I forgot to take photos of unfortunately, although it wouldn't have revealed much more really.

I took the PCB off to have a look at it. The actual power amp itself is "crude", but also quite clever in a way. It's purely an emitter-follower - just the two power transistors and a resistor arrangement under that potting compound, you can tell from the layout on the other side. The diodes and resistor on top are presumably for temperature feedback. Of course this means that it can't have more than unity gain, but for an "attenuator", this is exactly what you want! It also means that it is indeed transparent in terms of reproducing the signal it gets, although since that's from a dummy load rather than a speaker it won't have the same dynamic response as the unattenuated signal. (For what it's worth the "Plexi" switch on the back simply bypasses the inductive part of the dummy load which is in series with the resistors, further raising the true impedance and making it even more risky to the amp.)

The rest of the circuitry on the board is a relay which connects the output of the power amp to the speaker, driven by a turn-on circuit which senses input signal. From what I understand, this is to prevent any damage to the amp or the unit if you accidentally connect an amp to the output rather than the input - and this *is* a clever and well-thought-out solution, although it means you need to give the unit a bit of a blast before it will turn on at all, which is a bit disconcerting at first if you don't know that!

Although it's a trivial complaint really, I wasn't too impressed by the way the power transistors are mounted - long bolts (both bent!) which pass right through the heatsink and the casing. Functional enough but - like the crooked transistors - just not good workmanship. [/OCD] And again, no shakeproof washers anywhere so they can come loose. The rectifier is held on to the heatsink by a self-tapper, and had no thermal compound.

That's about it on the dismantling - I did finally take everything apart from the fan out though, because at this point I'd decided to fix it, and I wanted to mount everything properly.

ICBM

ToneControl said:

ICBM said:

ToneControl said:

can't you just trim the cone off a driver and use the voice coil as a load?

No, it will have almost no power handling capacity. The voice coil needs to be cooled by movement, without that it's just a coil of wire and will quickly burn out.

I thought some people ran them in a can of oil?

Yes, that would work.

SunDevil

So 'Ultimate' in this case is referring to it's potential to be the last thing you ever do - on that basis, you can't really fault them on the brand name?

Great info - terrifying little box!

ICBM

And finally… the rebuild.

As Danny said, of course four 30-ohm resistors can be wired all in parallel to give 7.5 ohms - which is almost exactly 8 ohms, and also close to the DC resistance of an 8-ohm speaker. So with the inductor on top as well, it should be about perfect as an 8-ohm dummy load.

I then worked out that if I set the four resistors at just the right angle, I could use both the original holes in the casing and fit them all in between the transformer and the side of the box, and just miss the power amp heatsink. Just! I couldn't get the spacing absolutely 100% equal, but close enough. I mounted them with thermal compound so the casing now becomes the heatsink.

I fitted a panel-mount fuse holder, cleaned up the mains cable hole and replaced the cable clamp with a larger one - and a UK-coded mains cable - and heatshrinked the connections. And fitted a proper earth terminal washer. (I know for the regulations, it needs to be a *separate* terminal, but I'm not going to be quite that strict since I won't be selling this.)



I replaced the speaker jacks with Switchcrafts, rewired the bypass switch to an 'x' pattern (both switch poles in parallel, almost totally reliable), and by re-arranging the wires from the pot and the inductor and adding one chassis-mount tag strip at the corner of the transformer I was able to eliminate all unsupported joints while not needing to replace any parts… quite a miracle! Replaced all the solid wire with stranded cable apart from the links between the resistors, which are totally rigid anyway.



All bolts, pots and jacks are fitted with shakeproof washers.

And because I'm not Mr. Perfect, I did bodge one little bit by accident . Actually while I was getting it apart, rather than rebuilding it - in my defence if I hadn't been trying to undo all that mess without damaging anything it wouldn't have happened.

I caught the corner of the relay with my soldering iron. Damn...

And I wouldn't actually say it's good enough to be sold commercially even now.

But the really interesting thing is this: it sounds much better now - very natural and much closer to the unattenuated sound! So not only is the 30-ohm load potentially dangerous to the amp, it's not even for a worthwhile reason. Daft.

I also have another idea - the main reason I can think of for re-amping as opposed to simple attenuation is to allow effects to be inserted *after* the power stage of a valve amp. To build a system like this but not allow that also seems a bit pointless, so I'm going to fit it with a loop. Given the high signal level it's likely it will need the same sort of step-down/up circuit as you need in a valve amp, but conveniently there is now a big free space above the power resistors where you could fit something like the Metro Amps kit… that will have to wait a while though.

So for the makers of this thing if for any reason they read this, I can only say: it's not the concept that's the problem. But for it to match the claims for it, it needs to have a more sensible load resistor value, and be built to a recognised standard of safety and good workmanship. It could do with a loop option (I thought that might have been offered, but it doesn't say so on their website), and probably with less hype and inaccurate advertising. Preferably before it blows up any more amps...

(Edit - old Photobucket pics replaced.)

ICBM

SunDevil said:

So 'Ultimate' in this case is referring to it's potential to be the last thing you ever do - on that basis, you can't really fault them on the brand name?

Great info - terrifying little box!

lol

ourmaninthenorth

Terrific write up ICBM. Thanks for posting.  

Sassafras

Your soldering certainly looks a hell of a lot better than the original wreck!

Even my soldering's better than that mess.

Rocker

Well done my friend.  Your work has made it look "right".  My instinct would be to fit flying leads for the input [from the amplifier] and the output [to the speakers] rather than depending on jacks and jack plugs.

Thanks for the brilliant write up.  And the photos.  There are times when a good photograph is worth a thousand words.  Your photos speak volumes, your words explain your thinking very well.

Years ago I came across a guitar combo amp, it was made in East Germany IMSMR, that had one of those inline fuse holders.  That combo was sent to the recycling depot in a skip that was hired when we were having our bathroom updated.  The speakers worked but were not worth keeping, nothing else was salvageable.

Drew_TNBD

ICBM said:

And because I'm not Mr. Perfect

Yes you are  :-*

Pete24v

Good lord!

I feel bad for even giving this away with the Cornell. In my defence, I paid £150 for it on this forum, and only used it twice at home, once with the Cornell and once with my Dual Showman Reverb.. And I didn't like the sound after a few minutes, so I disconnected it.

Sorry Lyndon, I hope the Plexi is ok.

Pete

ICBM

Pete24v said:

Good lord!

I feel bad for even giving this away with the Cornell. In my defence, I paid £150 for it on this forum, and only used it twice at home, once with the Cornell and once with my Dual Showman Reverb.. And I didn't like the sound after a few minutes, so I disconnected it.

Sorry Lyndon, I hope the Plexi is ok.

Pete

Not your fault Pete - nor the original buyer's. The company purports to be a professional manufacturer, and customers can't be expected to know about what's inside it. That's why there are safety regulations - which appear not to be a legal requirement in the US.

Luckily Fender built their amps well, so it didn't wreck your 4-ohm Dual Showman either.