New Hiwatt Project - 1977 DR505

[Raoul Duke]

So after keeping an eye out for a couple of years and watching the prices shoot into the stratosphere; I was recently lucky enough to strike a pretty reasonable deal for a 1977 Hiwatt DR505 that’s been mostly “un-Canadianed” by the previous owner. Before the holidays I’d moved some gear that wasn’t being used, so finances and opportunity seemed to converge serendipitously, lol. Having a quick play through and inspection here’s what I’ve found so far:

* This thing is super grungy! Might have been stored in a barn at some point as I’ve dusted out quite a bit of chaff. Also, there are corrosion spots that have actually started flaking off rust chunks here and there as well as signs of isolated moisture puddles (long since dried out) within the cab. Having built a few old cars into hot rods - I recognize this as the signs of mouse piss. Nothing serious - just plenty to clean.

* Powered up I got 486 vDC on the plates with -41vDC on pins 5 of the power tubes. As you see in the pics - mismatched 1R resistors of dubious nature were installed as bias sensors - I read ~60mV (mA) on these. Seems a little hot, eh? I got 6.4vAC across the filaments with 121vAC out of the wall - so I think overall operating voltages are within normal limits.

* Sounded pretty good. Controls work as I’d expect.

* Electrolytics have been replaced at some point. All 5 cans are TAD brand stock values with embossed 6 digit numbers in the flat bottom - all starting with “19”, i.e. “191005”. Anyone know if these are date codes?

My plan (as of right now anyway):
1. Install adjustable bias and get that set correctly.
2. Test and verify all caps and resistors.
3. Review the grounding scheme - I’ve read Hiwatts sometimes have loops built in?
4. Replace the goofy fuse holders, lol.
5. Verify all wiring and connections (mainly concerned with OT and PT).
6. CLEAN THIS THING!

Been studying Mark Huss’ awesome Hiwatt site for a few months now. Tremendous amount of helpful information there! These amps are certainly their own “thing”. Here’s some “before” pics:

IMG_0514.jpeg

IMG_0515.jpeg

IMG_0516.jpeg

IMG_0517.jpeg

I’ll update as I go, stay tuned…

[Raoul Duke]

Checked all the caps and resistors and everything but the treble cap on the preamp board were within spec. The treble cap is identified as 220pf but tested at 142pf - so I guess I need to decide whether to change it or leave it. I’m leaning toward changing it out since the aim here is to get the circuit working as it’s supposed to; but the purple ink on the solder joints makes me second guess myself, lol.

Started cleaning as well. Should keep me busy for a few days. Under the dust and grime on the outside of the chassis I’ve discovered a layer of oily residue. Also a ton of oily residue coming out of the pot shafts and all over the inner chassis wall where the pots are. Not sure what got sprayed in there, but it’s got to go.

[Raoul Duke]

Couple days later and I’ve done a lot of cleaning both to the chassis and cab. This thing definitely sat somewhere for awhile based on the fuzz, dirt, mouse turds, and other particulate matter that it was encrusted in both inside and out.

The outside of the chassis was coated in some type of oily residue that had absorbed decades of grime and hardened and bonded to the zinc coating. Getting under the transformers was difficult, so I loosened and tipped them (carefully) and cleaned underneath as well as re-seating a bell end grommet on the primary side of the OT that had partially pulled out. Removed the heat shield as the gap between it and the chassis was packed solid with gross chunks of effluvia.

The cab exterior got a brushing down/wiping off with some diluted Fantastic spray which completely browned a microfiber cloth as well as a good scrubbing inside with a tooth brush. All the hardware was removed and cleaned, vinyl on the handle was pulled back into center and re-seated, and new piping for the front panel cut to size and soaked in coffee overnight to reduce its brightness. I installed it without adhesive, but may re-do it with a few strategically placed dabs as it wants to pop out at certain points along its length. There was one rear panel cleat that had started to delaminate, so I soaked it with thin super glue and clamped it tight. Looks/feels/acts like plywood again. Probably give the cab one more going over while I wait for parts.

Started re-working the power and output sections as well. Rotated the cans so that they aligned better with the 90 degree bends in the wire runs, removed any charred insulation and added shrink tubing where necessary, and re-shaped many of the runs back to the HJ/Reeves standard. Even though it measured right on 1K, I replaced the can mounted 1W Iskra dropper with a 3W Vishay metal film. I see that modern build documentation calls for a 5W there, so hopefully I’m not taking a chance. I figured a modern 3W MF has to be safer than a 50 year old 1W CF.

Installed some Dale 1%/1R sensing resistors after cleaning the grounding lugs and chassis, and added a separate ground lug for the HT fuse, reservoir filter, and what I think is the heater CT as all these were wrapped around the V5 cathode ground lug. Trying to disturb as little purple ink as possible, but this amp is far from a museum piece and form has to follow function. Figuring out how to approach the adjustable bias as well. I have an idea for something in between Mark Huss’ “quick and dirty” and “improved” circuits. Basically it’s the former built on to a board with all the serviceable pieces easily accessible and removed from the PS/Output tag strip. I’ll post a DIYLC when I get a chance to make one up. Or I may just build the improved on a board. Still thinking it through.

Carefully unfastened and cleaned the shield ground mating surfaces for each pot with iso, then retightened. Also tried getting some date codes off the pots, but all I can see that isn’t self explanatory are a couple of digits on either side of the case notch. All the pots have the same digits: “2” on one side and “7” on the other (see pic).

IMG_0519.jpeg

IMG_0520.jpeg

Really enjoying these re-dos of old amps just as much as the builds. I’m finding it really interesting and a great learning experience; and the end results so far have been well worth the effort.

As always - any advice, info, or anecdotes are always welcome and appreciated!

[Raoul Duke]

Here’s what I was thinking about. It’s basically reinventing what’s already there, but moving it off the PS strip. I’d likely throw it up on a couple of stand-offs at the V8 tube knock out. Maybe it’s a lot of extra work vs just mounting the pot, but the pieces that would be removed for servicing have their own eyelets and it de-clutters the strip a little I guess.

DR504 Bias Board.PNG

[Raoul Duke]

So after thinking about it and reading the various threads I could find and of course studying the options on Hiwatt.org - I figured might as well put everything on the board. Here's my new idea:

DR504 Complete Bias Board.png

I built it using actual component dimensions just to see if everything fit in the space - and it seems it will. I ball-parked some of the values based on the various circuits I looked at and am prepared to adjust as needed.

Looking for constructive criticism just to be sure I haven't forgotten anything.

Thanks in advance!

[martin manning]

You may have to adjust the voltage range, but that looks good. Fail-safe and low ripple.

[Raoul Duke]

Thanks Martin!
I applied your fail safe design to this and the dual caps ala HAD.
I must be learning something from all the reading I do here, lol…

[Raoul Duke]

Some progress as I wait for parts coming tomorrow:

Re-did all the cap can connections while straightening out the wire runs and replaced the balance/bleeder resistors with 220k/2w Welwyns as well as replacing the screen grid resistors with 1K/5w Welwyns. Welwyns seemed like the right choice given the brand of amp, lol. I’ve read some suggestions that the standard 1K value should be upped to as much as 3K for today’s EL34s. Anyone have any thoughts on that?

Also prepped the PS/Bias strip for the separate bias board - all that’s left is to remove the diode, 110k dropper, and the 30k end stop. Got the stand-offs and the unpopulated board in to test fit; then cleaned and re-tightened every ground point - adding a couple modern ground lugs where necessary.

Replaced the IEC as it came with a wiggly ground prong/terminal. Also did some splicing at the mains switch due to some wire insulation that became some mouses’ dinner at some point. The wire had a huge wad of electrical tape on it, and when I peeled it off - there was absolutely no insulation left and half the copper was frayed. I don’t understand how that got overlooked during its previous service - but grafting on new wire isn’t complicated. I’d rather trust a soldered and shrink-tubed lineman’s splice than half the copper and electrical tape any day.

Also did quite a bit more cleaning both inside and outside the chassis and cab - not that it’s apparent…

IMG_0345.jpeg

[Raoul Duke]

You may have to adjust the voltage range, but that looks good. Fail-safe and low ripple.

So after re-reading some older bias circuit posts - most notably by RG Keen - here’s what I’m thinking:

If the 110k dropper from the HT running through the diode and filter cap with a 30k “end” was giving me -41v at pin 5 (raw bias voltage) it’s likely that the 110k will need to be lower so that I can get closer to -50v.

Doing the math with the component values on the drawing gets me .81 - which gives me roughly -33v.

If my math and application of what I read is correct…

[martin manning]

Simple ohm's law estimates are generally not accurate in this case due to the spiky nature of the current flow. Reducing the 110k is the right direction to get a more negative bias voltage, but you will likely be doing some cut and try.

[Raoul Duke]

Thanks, that’s what I figured Martin. Just the variance between matched tube pairs would seem enough of a variable.
I try to use the math wherever I can - even just for the practice - trying to see if it’s sinking in, lol.

[Raoul Duke]

Got the bias board finished and everything in - then did a first start (no tubes) on the current limiter.

* Wall voltage 121.2vAC
* Bulb behaved as it should
* Unloaded voltage readings-
** Heaters: 6.7
** Plate voltage: 548
** Bias pot sweeps from -27.7 to -55.6
* No strange sounds, smells, sights

I think this range should work well for two EL34s, so I put the tubes in.

Here’s the loaded voltages:

V1 plate A: 137 B: 222 cathode: 2.3
V2 plate A: 226 cathode: 1.5
V4 plate A: 349 plate B: 348 cathode: 80.5
V3 plate A: 216 plate B: 348 cathode A:1.9 B: 78.5
V4 plate: 509 screen: 507.6
V5 plate: 509 screen: 507.8

Some odd values based on what I’m used to - but this is my first Hiwatt. I’ll have to look for some comparable info for similar circuits. Biased the tubes at ~32 mA (they are well matched) at about 30% on the trimmer.

The bias board values are 100k dropping resistor, 15k load resistor, 18k in series with the trim pot. Also, my radial caps are 56uf/100v. I’m thinking if I go up on the 18k I’ll get closer to the center of the trim pot travel - but I’m not sure I need to. The sweep range as is goes from ~20 mA to 120mA. I’d think those values would bias a broad spectrum of tube specs?

Here’s a board pic before start-up:

IMG_0521.jpeg

What would yous guys do? Any advice is appreciated!

[Raoul Duke]

So I went ahead and replaced the 18k with a 30k and ended up with the same situation - but with the bias trimmer 180 degrees. Splitting the difference, I tried a 24k and now the adjustment is pretty much dead center at 60% tube dissipation.

Tried finding some voltage charts for this circuit and found that they aren’t common. I guess everything looks ok; my main question had to do with V1s drastic difference between triode plates. I swapped a new JJ in for the recent production Mullard it came with and had a similar disparity. Guess I’ll keep searching - but the sound test doesn’t indicate anything wrong and all the voltages are stable. The charts I was able to find were split pretty evenly between fairly even and disparate - and none were for a late 70’s two input 12AX7 PI circuit.

I let it run for about 90 minutes and everything stayed stable with no hums, hisses, or buzzes - so I guess I’ll put her back together.

[sluckey]

my main question had to do with V1s drastic difference between triode plates.

That voltage difference is because one triode has a 270K plate load resistor and the other triode has a 100K plate load resistor.

[Raoul Duke]

Ahh, now I see - thanks Steve!
I figured it must be a design thing since it followed two newer tubes - but wasn’t sure.

I always appreciate the help!