Selenium Bridge Rectifier On Geloso Tube Amp

[sarakisof]

Thank you Phoenix. So after X2 and 3 prong cable installed, full power recapping and replacing of  bridge rectifier all looked perfect. I was jamming today morning for about 3 hours with no problem.

But now returned home, powered it on and when Standby switch released on i get no sound, only pops in speaker. By evevery pop i get accordingly "pops" in fluorescent blue gas inside EL84s, they light on and off , pops accordingly.
What could this be? Was working like champ in the morning.
☹️
Should i power on without power tubes? Where i am starting of for troubleshooting?

[Rob Strand]

Check all the voltages including the voltages on the power tube cathodes.

It's also possible a tube has failed.

[sarakisof]

Check all those Rob. I think i found out. Bad old solder, cracked joint at output trafo's winding terminals.
I will actually replace all those Geloso's impedance outputs. OT give a few to 500 ohms output impedance so you select via those output taps.
What do you guys suggest as replacement output connectors for those old rusty taps?




EDIT: Could the above dropouts/pops be caused by that i am temporary using a 15W speaker i had on hand and amp is 30W rated?

[Phoenix]

Glad you found the problem.
If I were you I'd just try to clean up those screw terminals, maybe replace the screws themselves if they don't clean up too well, but I'd leave the terminal block in place.

Using an underpowered speaker is risky - make sure you don't fall into the trap of thinking that half-way up on the volume dial is half power, remember that volume/gain controls are not *power* controls, just voltage controls. What exact voltage the amp will put out at any given setting is arbirary, there'll usually be some excess gain (a setting beyond which the signal doesn't increase in peak amplitude, only begin clipping, increasing average amplitude instead), and depends on input signal voltage and headroom of the various stages of the pre and power amp.
Always a good idea to use a speaker(s) with greater power rating than the amp it's(they're) connected to, as any clipping can increase power output beyond the clean "nominal" rating of the amp. A factor of at least x1.5 is a good idea, as a heavily clipping amp will often be able to sustain those levels when distorting, so a 45W speaker is a good idea.

It would be rare for a speaker to cause dropouts or pops, they tend to exhibit other failures if they're overloaded, like a burning smell or just going open circuit.

[ThermionicScott]

Check all those Rob. I think i found out. Bad old solder, cracked joint at output trafo's winding terminals.
I will actually replace all those Geloso's impedance outputs. OT give a few to 500 ohms output impedance so you select via those output taps.
What do you guys suggest as replacement output connectors for those old rusty taps?



EDIT: Could the above dropouts/pops be caused by that i am temporary using a 15W speaker i had on hand and amp is 30W rated?

My Italian isn't too good, but is there a reason for the wire between terminals 1 and 2?  If you had an 8-ohm speaker for example, I would it to be connected between terminals 1 and 5, with no other terminals used.  Looks like this amp could be used for line (PA/Muzak) applications too.

[Phoenix]

My Italian isn't too good, but is there a reason for the wire between terminals 1 and 2?  If you had an 8-ohm speaker for example, I would it to be connected between terminals 1 and 5, with no other terminals used.  Looks like this amp could be used for line (PA/Muzak) applications too.

The secondary of this transformer has a number of coils (appears to be 3 without doing any working out). A minimum of 2 of them must be used in a series or parallel combination for the rated power output (much like a power transformer with dual windings, say 15V for example's sake. You can get 15V out of one, but only half the current, to get the full power out of the transformer they need to be used in parallel (for double the current) or series (for double the voltage).

First column is the terminals the speaker attaches to, second and third columns are connections between windings. The various terminations provide a range of impedances. This type of arrangement was quite common on "universal/radio replacement" output transformers of old.

[ThermionicScott]

Aha, thanks, Phoenix.

[PRR]

.....common on "universal/radio replacement" output transformers of old.

Also on the Williamson super-spiffy wideband amplifier. As you say, several/many secondaries interleaved with many primary sections for good coupling. Then connect most or all secondaries in series/parallel combinations to hit your desired load impedance.

[sarakisof]

Well this will be a long post, my apologies.

Glad you found the problem.
If I were you I'd just try to clean up those screw terminals, maybe replace the screws themselves if they don't clean up too well, but I'd leave the terminal block in place.

Me too Phoenix (and i have already cleaned as much as i can), but unfortunately in my situation amp found with a mess back there already. One screw connector is completely missing (solder- hardwired) and most of the others are loose.



I can live with that and replace cables with new ones only (in the pic there are the cables it came with 🤣), doing some last more stable connections mounting/bonding etc.
But i have a silly impression from childhood (papa n grandpa used those screw terminals for listening all the time you know 🤣) that those screw connectors/copper wire aren't the ideal way to go 😁. That's why i asked if banana plugs, jacks etc. could be better.
I was thinking of placing new plugs for each screw terminal in their old place without removing the old block. I hate replacing such old beauties in love and respectively for their era. And as i grow up i realise that even for components i see myself replacing only the faults. Not exactly "if it isn't broken, leave it" but you get an idea. An old e-Ducati cap measuring some μf more isn't a big deal for me nowadays, in early years i was replacing everything.
If you think something else than those screws would make a better connection there i will go for it. In other case i will leave it.

Here is what i have done so far guys.

3 prong cable, X2 cap, bridge rectifier, electro caps and some polyester that went really bad.














Results:
With 0V volt selector i get 342Vdc at B+ and 6.10Vac at power heaters.
With +10V volt selector i get 328Vdc at B+ and 5.70Vac at power heaters.
Which option should i go with? (Will not go for drop resistor.) Original calls for 320Vdc B+.
Or go with -10V volt selector option so heaters get almost 6.3 but with cost in B+ more bigger?.

It plays great, this old thing has a lot of beauty in its soul. ✌️
I also spent a few days searching n working on chassis restoration. Realised that the best way to go is with sandblasting and then powder coating. Have no access on those things and they cost, so after searching for matching same hammertone effect realised i cannot get the same results with original. So i just used my favorite Fertan rust converter solution (not remover, uses tannic acid, non abrasive) which converted the old rust to a nice black grey film when srink with a lil water damo cloth. And then i just sprayed it with a metal varnish mat can and that's it. The beast shows his age without brown red rust on.

Next days i will see if i do mods for guitar 68K / bass inputs.

[Phoenix]

Me too Phoenix (and i have already cleaned as much as i can), but unfortunately in my situation amp found with a mess back there already. One screw connector is completely missing (solder- hardwired) and most of the others are loose.

Ah, your last photo didn't show the damage to the phenolic board. If I were you I'd be tempted to fashion a new board from FR4/G10 epoxy fiberglass (turretboard material) and use turrets in place of the screw terminals. It'd be in keeping with the style, and keep the functionality while being nice and reliable. The impedance options are handy, but probably not something you'd need to change often enough that a soldered connection would be much hassle.

But i have a silly impression from childhood (papa n grandpa used those screw terminals for listening all the time you know 🤣) that those screw connectors/copper wire aren't the ideal way to go 😁. That's why i asked if banana plugs, jacks etc. could be better.

I've found that style of screw terminal to be pretty reliable other than the phenolic mount going brittle like that, but yes, good idea to replace it given it's broken.

With 0V volt selector i get 342Vdc at B+ and 6.10Vac at power heaters.
With +10V volt selector i get 328Vdc at B+ and 5.70Vac at power heaters.
Which option should i go with? (Will not go for drop resistor.) Original calls for 320Vdc B+.

*If* your wall voltage was nominal when you made those measurements (ie, not particularly high or low. My voltage, nominal 240V, wanders anywhere between ~235V and ~255V-ish over the course of the day), then 0V selector is the way to go. Targeting as close to 6.3V as you can allows the amp to function correctly if the voltage is a bit high or a bit low. Heaters are +/-10% tolerant, so as long as it stays within the range of 5.67 and 6.93 within your wall voltage tolerance (conveniently also +/-10% in many countries), everything will be fine.

Everything else is looking ok, but I'd prefer to see those electrolytic caps have their leads mounted to something like tagstrip rather than floating like that. I know they're zip-tied down (holes in the chassis, or self adhesive cable tie mounts? I wouldn't trust self-adhesive long term), but I still prefer not to leave thing waving in the breeze like that.

[antonis]

Next days i will see if i do mods for guitar 68K / bass inputs.

I'll keep an eye on 24/7 beaking news from Giannena, focusing on blazes... 

[sarakisof]

I know they're zip-tied down (holes in the chassis, or self adhesive cable tie mounts? I wouldn't trust self-adhesive long term), but I still prefer not to leave thing waving in the breeze like that.

Had no eyelet terminals in hand. They are zip-tied and stuck in place with glue gun adhesive. They are solid rock stable, i had always nice results with this method so far.

Next days i will see if i do mods for guitar 68K / bass inputs.

I'll keep an eye on 24/7 beaking news from Giannena, focusing on blazes... 

Χαχα που θα βρούμε κόσσες φτηνά με το κιλό ρε συ Αντώνη? 

[antonis]

Χαχα που θα βρούμε κόσσες φτηνά με το κιλό ρε συ Αντώνη? 

7.30€ with spare blade..
(the Grim reaper is warranty included..) 

[sarakisof]

7.30€ with spare blade..
(the Grim reaper is warranty included..) 

Heh we are trying to do serious business here Antonis and you provoke the discussion. 

So, just got measurements for bias.
Here we go



Plate To Cathode Voltage: 322Vdc
PC: 69.8A / 63.8mA
PD: 22.48W / 20.55W

And think that cathode resistor measured about 130Ω instead of 160Ω. (That's cool, because if 160Ω i would had more cool bias  )

So for EL34 25W i think i am about 90% and 83%. A bit cool but as amp sounds sweet i think i will leave it as is. What do you think?

[Phoenix]

Plate To Cathode Voltage: 322Vdc
PC: 69.8A / 63.8mA
PD: 22.48W / 20.55W

And think that cathode resistor measured about 130Ω instead of 160Ω.

So for EL34 25W i think i am about 90% and 83%. A bit cool but as amp sounds sweet i think i will leave it as is. What do you think?

Seems reasonable, not sure why you think it's on the cool side, the target isn't 100% dissipation at idle for most push-pull amps.
If you're planning on overdriving the power amp much I'd be tempted to cool it off a little, get that cathode resistor back up to the original 160ohms, but you'll probably get away with leaving it be. Just keep an eye out for redplating when you give it a proper full volume test.

[Rob Strand]

And think that cathode resistor measured about 130Ω instead of 160Ω. (That's cool, because if 160Ω i would had more cool bias  )

There's actually a 1k in parallel with the 160 ohms,
1k // 160 = 137 ohms

Cathode on the schem is 19V.
19V / 137 / 2 = 69mA per tube cathode current.

Close to your values.

[Phoenix]

There's actually a 1k in parallel with the 160 ohms,
1k // 160 = 137 ohms

Whoops, missed that!
As Rob says, very close to schematic values (which should be considered to be within +/-20% at best anyway).

Only thing to watch for is redplating, as this wouldn't have been designed with overdrive in mind. If it does, cool it off a little, no harm in that.

[sarakisof]

Oops i missed that too.
Yeap i have 19-20V on cathode. Great. 
And if redplating in future (don't think so as i play clean
mostly) i could raise the cathode resistor as Phoenix said.
Perfect.

[Rob Strand]

Yeap i have 19-20V on cathode. Great. 
And if redplating in future (don't think so as i play clean
mostly) i could raise the cathode resistor as Phoenix said.
Perfect.

It occurred to me later the 19V cathode resistor drop reduces the plate voltage a bit and reduces the plate dissipation.

VPK = 320 - 19 - 80*69mA = 320 - 19 - 5.5 = 295.5V
@295.5V Ibias = 69mA is 20.4W
82 % Pdis_max of 25W for the EL34

70% Pdis_max is about 59mA

My understanding that amp is a HiFi amp so they might have bumped the bias up a bit to shave some distortion off.

Sorry I thought your 320V was the power supply but it's actually VP - VK.

[sarakisof]

Sorry I thought your 320V was the power supply but it's actually VP - VK.

Yeap