Carlsbro 50/Express hybrid/monster!

[Structo]

greatwhitemark,
Do you own a variac?

I was thinking to test it at a lower supply voltage and if it improves quite a bit, sell that PT and get one with the correct voltage.
Or hang on to it so you can revert the amp back to a Carlsbro if you decide to sell it.
Not familiar with those so I don't know what kind of money they go for.

[Cliff Schecht]

Zener's are extraordinarily noisy, especially when they are near the avalanche voltage. You have to be careful to put them in places where they can't contribute a significant amount of noise and like Paul says, you need to properly bypass them as well.

The awesome thing about Zener noise is that it is VERY broadband. Heck, I use a Zener diode as the noise source in the new PAiA Noise Source if that tells you anything.

[Structo]

A

I was looking at your schematic.

Do you already have the 245v primary connected?

If not, that will lower the voltages a bit if you use that tap.

I used a 25 watt Zener on an amp, I'm always a little gun shy to push that one too hard, but it's still hanging in there.

As for the noise, I just hung a .01uF ceramic on the Zener and it's quiet.
I don't hear any switching noise at all on that amp.
Especially if I play a single coil guitar through it.

[greatwhitemark]

Just got back from rehearsal room and managed to have a blast with the ‘Carlspress’ and must say that I really like it!
I’m fortunate enough to have a Ron Worley built clone at the room which is spec’d pretty close to an original so I could do a direct comparison. The Carlspress was slightly less focused in the bottom end and felt a bit spongier but not in a bad way, it certainly held together much better than the original Carlsbro or any Marshall that I’ve tried. Ron’s clone definitely had the focus. I’m basically well chuffed with the way it’s turned out, so far....
There’s still the issue of the high voltage and what effect it could have on components and stability. My immediate fear is it eating up tubes, but would like to know of any other potential problems?

ampgeek – would be great to see that spreadsheet if poss, and would be interested to hear any ideas for dropping the B+. Is it possible to get things to a safer level with dropping resistors?

Thanks everyone for sharing your thoughts on the zener/cap combo. It’s good to know that it’s been successful. Is it better to use a single zener or a string of smaller ones? I’m a little reluctant to go for it with the noise and reliability issues. As I said, I’m stoked with the way the amp sounds but just a bit concerned with the voltage and potential damage to tubes and components. Any other thoughts greatly appreciated....

[ampgeek]

NICE!! Makes you feel good..doesn't it?!?

If you conced the fact that you are not going for a nuts on clone, I would do the following. Mind you, I am definitely not an experiencee EL34 guy nor a Trainwreck expert by any stretch of the imagination. But...I am fairly adroit at reading the tube charts...and..asking the right questions.

Put 1 ohm precision resistors between EL34 pins 1 & 8 and ground for measuring current if you haven't already.

With 500 V at the EL-34 plates, bias to 35 mA at the plates. Look for ~39 mV across the 1 ohm cathode resistors as there is a current contribution from the screen grids. Nominally, 10% of the total.

Tweak the bias (within reason) to taste and measure and record B+ voltages for reference. Many folks report a better tone/feel with slightly hotter bias.

If B+2 is still higher than you are comfortable with, change the value of R25. I poked around a little and found that a number of folks run EL34s in the 500 V range and the screens in the 425 V range. Again, approximate values here.

I calculate that a 5K value at R25 will drop the screens to ~425 V and put the rest of the B+ string, as is, really close to the Express reference values (which vary a bit across the documents in the TW files sticky thread). If you are going to buy parts, pick up a range of values between 4K and 6K so you can tweak as you see fit.

I am away from home (and my tech documents!) for a week or more so I can't forward you my "model". But...I downloaded the Express voltage chart and did some calcs there. See attached in blue field.

The "Baseline" calcs simply calculate the current flow through each tube based on the voltage drop through Rk. I had to make an assumption for EL-34 screen grid current as you will see.

Knowing those, you can calculate (with Ohms Law) voltage drops across known resistor values.

If someone knows the voltage drop across the screen resistors....we could eliminate that assumption and lend more confidence to all of this.

Good luck,
Dave O.

EDIT
Arghhhh....I found an error when I proofed it after posting. Forgot to add in V3 current contribution. So...with that....I would suggest changing R26 a&b to ~8k4 (or whatever the closest standard value to that is) to get closer to "reference" values on down the line. Attachment updated.

[greatwhitemark]

Thanks Dave, that really is awesome! And yes it does feel good...........it was never about creating a clone as such, just using what was available.
I need to get hold of a few bits before I can try these suggestions.
Can the 1 ohm resistor be left in place? Or does it need to be removed and replaced every bias change? What power rating?
Those 25w resistors are hard to come by in uk, as is the 1 ohm precision.

Will let you know how I get on.

[ampgeek]

Right on Mark! Glad to be able to help out.

I buy LOTs of stuff from Mouser here in the US and I hear that some of our international bretheren here do as well. I note that they do have a UK site that might make your life tons easier.

http://uk.mouser.com/Home.aspx

There is absolutely no practical/operational problem in leaving the 1 ohm installed. I do that in all of the amps that I build. Makes life WAY easy when it comes to biasing up the power tubes. The Express Bill of Materials lists those as 2 watters.

R25 at 25 watts provides a HUGE margin over what one would calculate the requirement to be (like...1.5'ish watts!). Not sure why and hopefully someone will chime in on this. But...I would also look at 10W options without concern. Overall physical length may be an issue with the lower watt options (they are shorter). But..I have never had problems with heat shrunk splices.

I recommend pulling up the Express bill of materials, opening up the UK Mouser page and start cutting/pasting in the P/N from the BOM to see what is available. Follow links to their catalog pages to see what range options are available. I use their Project Manager function to easily keep track of what I need for a given project. A great feature if, like me, you tend to chip away at these things little by little over a fairly long time period!

Cheers,
Dave O.

[Structo]

You can leave the 1 ohm resistors in place.
The critical thing is to get precision resistors with less than 1% tolerance.

[greatwhitemark]

Mouser is the only place that I've seen the 25w cement type resistors, the down side is the shipping costs at £12 for an 0.82p component. I thought I could make it worthwhile by fleshing out the order with a few other things that I need, but of course they didn't stock them at Mouser - they stock them at Digikey, another US company with a £12 shipping charge. So would be almost $40 shipping for a few dollars of components.

But I know my local Maplin does 10w cement type, so thanks for the heads up. Would 2x10k 10watt resistors in parallel make a 5k rated at 20w? Purely theoretical as the highest value is 4k7 in Maplin’s 10w range anyway. I’ll give one a try!

Tom(Structo) it’s a similar situation with the precision resistors, I can find a few with great tolerance but the values don’t go as low as 1ohm. I think the best I can find is 1% so will have to go with that for now.

Onwards and upwards........

[paulster]

Mouser is the only place that I've seen the 25w cement type resistors, the down side is the shipping costs at £12 for an 0.82p component. I thought I could make it worthwhile by fleshing out the order with a few other things that I need, but of course they didn't stock them at Mouser - they stock them at Digikey, another US company with a £12 shipping charge. So would be almost $40 shipping for a few dollars of components.

Yes, really annoying. I've ended up having to place large orders with Mouser to make the shipping worthwhile. If Mouser did all the PVC series caps it would make life a lot easier!

But I know my local Maplin does 10w cement type, so thanks for the heads up. Would 2x10k 10watt resistors in parallel make a 5k rated at 20w? Purely theoretical as the highest value is 4k7 in Maplin’s 10w range anyway. I’ll give one a try!

Yes, that would work. Have a look at Farnell, CPC, Rapid Online and RS for the ceramic resistors as none of these require you to have a trade account these days for online orders.

Tom(Structo) it’s a similar situation with the precision resistors, I can find a few with great tolerance but the values don’t go as low as 1ohm. I think the best I can find is 1% so will have to go with that for now.

Check out http://www.hificollective.co.uk and find the 1W Holco H2 1ohm resistors. They're not cheap but they're 0.5% tolerance, so perfect for this application and it's not like you need a load of them.

[Cliff Schecht]

If you know the resistance of the resistor then it's not hard to calculate your corresponding bias voltage. I use 5-10% parts for my bias measuring resistors because they're what I have on hand. Just measure the resistance, mark it down, divide the voltage by this resistance and you have the current.

[paulster]

If you know the resistance of the resistor then it's not hard to calculate your corresponding bias voltage. I use 5-10% parts for my bias measuring resistors because they're what I have on hand. Just measure the resistance, mark it down, divide the voltage by this resistance and you have the current.

Very true, but I find it's a lot easier to know that they're actually 1ohm rather than a nominal value for each tube socket that you then have to factor in. Especially on 100w builds where I want to be able to quickly move between all the power tubes checking the bias whilst adjusting it.

There aren't many meters (save for the 4-wire ones) that can accurately measure resistance down at the 1ohm level, so I find for this particular part it's just easier overall to buy a very close tolerance part.

[greatwhitemark]

Check out http://www.hificollective.co.uk and find the 1W Holco H2 1ohm resistors. They're not cheap but they're 0.5% tolerance, so perfect for this application and it's not like you need a load of them.

I've already looked at hifi collective as they have a nice range of quality resistors, but when you click to buy they don't have all values in stock. The Holco starts at 12R.

I'm willing to go with the 1% option, buy a few and test for best. That's also a good suggestion to measure the actual value and calculate using that. Although, I'm a little nervous of accuracy coming along and ruining my build

[Rob S]

So THIS is what keeps a Young Man off the streets then?

Sounds as though you're having a ball Mark. I haven't forgotten the email I owe you, just been busy, busy, busy.

Rob

[greatwhitemark]

Hi Rob, yeah I've been in full mad scientist mode the last few days..........loving it!

Note to self: Eat and wash occasionally.