Orange or120 glow screen grid

[Chuck H]

Sounds like a leaky coupling cap to me. Replace C23 and C24.

Interesting. And worth looking into. If the amp is already taxing the screens and the bias is moved they may well start glowing before the plates. Of course the amp is still punishing the screens and that should be dealt with too. But if it was working before, and now it doesn't, and all these other possibilities have been covered, I say sub in some new caps and see. Nothing to lose.

[binito]

Sounds like a leaky coupling cap to me. Replace C23 and C24.

This caps are new. In the schematic I marked in green i changed

[Mick Bailey]

The positioning of the additional resistor is fine along with leaving the choke where it is. I recently had a Roost amp with very high screen and plate voltages and the main thing was to get the screens down from 560v. This uses a similar resistor arrangement (though the Roost has a 470 ohm resistor feeding individual 1k screen resistors). I ended up with 2.2k screens and quite a large main dropper resistor of I think 6.8K.

The problem with comparatively large resistors in the screen supplies is the voltage drop is current-dependent. I find that getting the quiescent screen voltage correct means the drop is too much at higher signal levels and can result in too much compression and loss of pick attack, volume and dynamics. This can be slightly improved by adding an additional cap (or more conveniently two caps in series with balancing resistors) to ground after the main dropper resistor to form a supply node.

Another approach is to use a voltage divider with a fairly robust dropper resistor and a lower-rated bleeder resistor. This works well, but is inefficient due to the constant drain across the supply. It also generates additional heat.

By far the best method I've found for dropping HT is to use a MOSFET, as per J M Fahey's suggestion.

[g1]

But the c5&c6 (100uF now) Yes, I'm going to change by a 33uF
By be 100uf c5 &c6 could be the problem, although the voltage is so high ?

Even if it is not the cause of the problem, it is probably making it worse. Changing back to the stock value should reduce the screen glow issue. If it doesn't cure the problem, then you can try a resistor in series with the choke.
But I'm very curious to know how much difference the proper cap values will make, before adding any resistors.

And another thing to consider, instead of dropping voltage with a resistor in series with the choke, you could add a bucking transformer to the primary (see R.G. Keen vintage voltage adapter Vintage Voltage Adapter). This will lower ALL the voltages, including the heaters.

[binito]

The positioning of the additional resistor is fine along with leaving the choke where it is. I recently had a Roost amp with very high screen and plate voltages and the main thing was to get the screens down from 560v. This uses a similar resistor arrangement (though the Roost has a 470 ohm resistor feeding individual 1k screen resistors). I ended up with 2.2k screens and quite a large main dropper resistor of I think 6.8K.

The problem with comparatively large resistors in the screen supplies is the voltage drop is current-dependent. I find that getting the quiescent screen voltage correct means the drop is too much at higher signal levels and can result in too much compression and loss of pick attack, volume and dynamics. This can be slightly improved by adding an additional cap (or more conveniently two caps in series with balancing resistors) to ground after the main dropper resistor to form a supply node.

Another approach is to use a voltage divider with a fairly robust dropper resistor and a lower-rated bleeder resistor. This works well, but is inefficient due to the constant drain across the supply. It also generates additional heat.

By far the best method I've found for dropping HT is to use a MOSFET, as per J M Fahey's suggestion.

the application with the mosfet is interesting. where I can pull a sample schema to apply to this amp

[nickb]

the application with the mosfet is interesting. where I can pull a sample schema to apply to this amp

If you are just dropping the screen voltage a single 5W 39V zener @ $0.34 might be the simplest option 1N5366BRLG ON Semiconductor | Mouser placed in series with the feed to the two screen resistors.

[Mick Bailey]

the application with the mosfet is interesting. where I can pull a sample schema to apply to this amp

Take a look at R.G's Mosfet Follies.

Also, scroll down to the bottom of this article for an example of how I build the circuit;

http://www.audiotesla.org/6146B%20Tweet.htm

It would be a good idea to measure the preamp plate voltages, too. You may be better off dropping both this and the screens, or even considering dropping the HT supply overall to give a B+ of perhaps 480v loaded. Then see what happens to your screen voltages and perhaps increase the screen resistors to 2.2k (you'd then end up with a similar situation to a Z-series JCM900)

[binito]

The positioning of the additional resistor is fine along with leaving the choke where it is. I recently had a Roost amp with very high screen and plate voltages and the main thing was to get the screens down from 560v. This uses a similar resistor arrangement (though the Roost has a 470 ohm resistor feeding individual 1k screen resistors). I ended up with 2.2k screens and quite a large main dropper resistor of I think 6.8K.

The problem with comparatively large resistors in the screen supplies is the voltage drop is current-dependent. I find that getting the quiescent screen voltage correct means the drop is too much at higher signal levels and can result in too much compression and loss of pick attack, volume and dynamics. This can be slightly improved by adding an additional cap (or more conveniently two caps in series with balancing resistors) to ground after the main dropper resistor to form a supply node.

Another approach is to use a voltage divider with a fairly robust dropper resistor and a lower-rated bleeder resistor. This works well, but is inefficient due to the constant drain across the supply. It also generates additional heat.

By far the best method I've found for dropping HT is to use a MOSFET, as per J M Fahey's suggestion.

Can i put a zener diode in series with B+ to dropping voltaje??
I attached a example of schematic:

https://dl.dropboxusercontent.com/u/...R120_zener.jpg
1N3335B - SOLID STATE - DIODO, ZENER, 5%, 62V, 50W, DO5 | Farnell Espaņa

It is sufficient to reduce b+? Or i need more things?
Thanks in advance

[nickb]

Can i put a zener diode in series with B+ to dropping voltaje??
I attached a example of schematic:

https://dl.dropboxusercontent.com/u/...R120_zener.jpg
1N3335B - SOLID STATE - DIODO, ZENER, 5%, 62V, 50W, DO5 | Farnell Espaņa

It is sufficient to reduce b+? Or i need more things?
Thanks in advance

The zener would only be practical for the screens. If you want to drop the HT you'll have to use the MOSFET dropper because of the higher power dissipation. http://www.geofex.com/Article_Folder...osfetfolly.htm

RG does mention a heat sink but I'd like to stress this - it's essential to have one to get rid of all the heat generated by the MOSFET.

[xjbear]

Theres room for 6550's and the current to run them. Midern el34 cant live in that environment long. In the US ww can rewire the 120 volt of the pt and neutral the othee leg to get the or120's down to about 485v but your dilemma will be a power scaling kit or diff tubes.

[Mick Bailey]

I usually bolt it to a cleaned area of chassis with some heatsink compound. Depending on the MOSFET, some have an isolated package that can be bolted onto an existing hole, others will need an insulated mounting.

[TimmyP1955]

This caps are new. In the schematic I marked in green i changed

A new one could be bad. Given that the problem changed sides when you swapped the outputs, it makes sense.

[Chuck H]

A good trick for the heat problem created by high power zeners is to parallel a big resistor valued to drop the intended voltage. Then the parallel zener just picks up the slack as it regulates. The zener can be downsized a little and you probably won't need to heat sink. Since resistors seem to handle power dissipation and heat with more aplomb you don't usually get those pesky self desoldering problems either. The downside is that it takes two components instead of one. Of course a heat sink is usually more expensive than a resistor.

[nsubulysses]

The positioning of the additional resistor is fine along with leaving the choke where it is. I recently had a Roost amp with very high screen and plate voltages and the main thing was to get the screens down from 560v. This uses a similar resistor arrangement (though the Roost has a 470 ohm resistor feeding individual 1k screen resistors). I ended up with 2.2k screens and quite a large main dropper resistor of I think 6.8K.

The problem with comparatively large resistors in the screen supplies is the voltage drop is current-dependent. I find that getting the quiescent screen voltage correct means the drop is too much at higher signal levels and can result in too much compression and loss of pick attack, volume and dynamics. This can be slightly improved by adding an additional cap (or more conveniently two caps in series with balancing resistors) to ground after the main dropper resistor to form a supply node.

If you are just dropping the screen voltage a single 5W 39V zener @ $0.34 might be the simplest option 1N5366BRLG ON Semiconductor | Mouser placed in series with the feed to the two screen resistors.

I have a question regarding these two posts.

I have messed around with amps that have high screen voltage that is too taxing on the screen, so the dropping resistor in series with screen supply and screen grid resistors was used to make the screen voltage drop at high volumes so the screen dissipation does not exceed the tubes max dissipation rating too badly. If I were to use a zener to drop screen voltage as Nick B suggests, rather than using a dropping resistor, I would assume I would basically just have a screen voltage that is lower but varies much less.

If I were to use the zener method would I be able to get the same power output but with less compression (because screen voltage varies less)? Is this an accurate assumption if anyone has messed around with both scenarios. It seems like it would make the amp sound better at high volume vs. the dropping resistor method. Just wondering before I order some zeners and do my own test.

[J M Fahey]

If I were to use a zener to drop screen voltage as Nick B suggests, rather than using a dropping resistor, I would assume I would basically just have a screen voltage that is lower but varies much less.

If I were to use the zener method would I be able to get the same power output but with less compression (because screen voltage varies less)? .

It will work the opposite way.
You are NOT stabilizing screen voltage because you are not referencing your Zener to ground but itīs full floating.

As in: your screen supply drops, say, 22 volts ... screen voltage will drop exact same 22 volts.
No stabilization.

Now if you build a MosFet dropper supply, whatever is at the Mosfet gate will be at the screens (shifted up some 3.5V).

IF mosfet gate comes from a voltage divider (fixed or a potentiometer) it will be a fixed percentage of whateverīs available at +V and essentially work like in an amp where screens are connected the old way, it will behave like we traditionally expect.

IF Gate voltage is stabilized, it will act as if amp had a very stiff power supply.

You choose what you prefer.