• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Metal Oxide Screen Stoppers

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I read somewhere on the internet that Metal Oxide(not metal film) resistors are similar in manufacture to Carbon Composition resistors just that the resistive material is different. Can't find the reference right now, but I'll keep looking.

I was looking for Carbon Comps for screen stopper duty, but they seem to be harder to find (or ridiculously expensive) at power ratings over 1/2 watt new. Metal oxides are much cheaper and available in higher ratings than carbon comps. I don't really want to use a dingy old one, either. Would a Metal Oxide have similarly low inductance and work as well in this application?
 
Still, if the metal oxides are non-inductive like the carbon comps, they are much cheaper and much higher power ratings are available.

I wanted to purchase a selection for experimentation, thus the desire to buy the cheaper metal oxides. If they work as well as carbon comps, I'd rather go that route. Do they? Anyone know of anyone who has done inductance measurements on these?

I guess I could always buy comparable parts and measure the inductance myself.
 
Administrator
Joined 2004
Paid Member
Metal Oxide resistors have pretty good surge ratings. Choose your power rating for it's maximum voltage rating. I haven't used Metal Oxides types for screen resistors, but I'll bet they work fine.

BTW, they don't really burn. They smoke a bit, then go open. No flames from what I have seen so far.

The way they are constructed is much like a carbon film resistor. They are deposited on a ceramic (or similar) core. Then they are trimmed using a helical cut. Compared to the resistance, I wouldn't call these inductive.

I guess I could always buy comparable parts and measure the inductance myself.
I can and have done so. They were unremarkable in that regard. That means that there was nothing I needed to remember (not a problem). I can always measure again if needed.

-Chris
 
I have been using metal oxide resistors as screen stoppers in many of my amps including the Simple SE. If you chip the coating off of some of these resistors you find that they have been laser trimmed in a spiral pattern. This has not been an issue with any of the common power tubes that I have tried.

I thought the big problem with film resistors on the screen grids was that the resistors could catch fire and burn.

YES! I found this out early in the SimpleSE design. Most experts say to use a 1/2 watt screen stopper. After I blew one and set another on fire, I will say to use at least a 1 watt, maybe 2 watts especially with an EL34 that gets pushed to clipping. Of course I tend to "test" things a bit more strenuously than the average user does. All of my amplifiers must be able to live with me playing my guitar through it with a guitar preamp set on kill. That takes out the wimpy designs, including the two previously mentioned screen resistors.

Beware some "carbon comps" aren't what they look like. Crush one with vice grips and see whats inside. If it's full of black charcoal looking stuf that crumbles. they are real. I have seen some that have a carbon film resistor inside!
 
http://www.koaspeer.com/pdfs/res52.pdf

The above data sheet shows the type of resistor I had in mind. Unfortunately, the picture seems to imply that there is some helical cutting (which I am assuming would increase inductance).

The data sheet also implies some flame resistant properties.

Unless a resistor guru weighs in and clears up all of my confusion, I may just have figure this out the old fashioned way.

edit: I'm a slow poster, it seems.
 
Administrator
Joined 2004
Paid Member
Hi SpreadSpectrum,
Okay so to sum it up: carbon comps are still best, but metal oxides are probably good enough.
I wouldn't say so. Carbon composition resistors are best in high frequency RF work. With the frequencies you are working with, the metal oxide parts are not considered inductive.

Hi tubelab,
I have been using metal oxide resistors as screen stoppers
Me too, on rebuilds and some playing around.

Most experts say to use a 1/2 watt screen stopper. After I blew one and set another on fire, I will say to use at least a 1 watt, maybe 2 watts especially with an EL34 that gets pushed to clipping.
The voltage rating on metal oxide types may have been your problem there. I often use 2 watt parts in tube amps for the voltage breakdown rating alone. Some 1/2 watt parts may only be good to around 200 VDC from memory, maybe 250 VDC? If it arcs over, you will see some cool effects!

-Chris
 
Carbon composition resistors are best in high frequency RF work. With the frequencies you are working with, the metal oxide parts are not considered inductive.

That's what I was trying to say, just didn't say it very well. I meant 'best' in the obsessive-compulsive sense.

I fully agree that metal oxides are good enough for screen stopper duty in an AF amp, but what about a replacement for higher power carbon comps in RF circuits? Would inductance in ceramic composition resistors be as low as carbon comps? The Ohmite data sheet indicates that they are carbon composition replacements. I don't know if that means in every sense.

http://www.ohmite.com/catalog/pdf/ox_oy_series.pdf
 
The voltage rating on metal oxide types may have been your problem there.

Ordinarilly the voltage rating covers the voltage rating voltage from the resistor to its surroundings, and voltage from end to end. How would the resistor know the absolute voltage from the resistor to ground if it is suspended in free air, or mounted on a PC board with no nearby traces, which is the case for the Simple SE.

The EL34 draws a lot more instantaneous screen current at the moment of "this one goes to eleven" style overload than we think. Under extreme conditions the plate voltage can be less than the screen voltage for a brief instant. The screen can draw considerable current for this time. The average current drawn by the screen grid may be 10 mA leading to a 100 milliwatts of dissipation in a 1K screen resistor. The peak current can hit 100 mA which will burn 1 watt in that resistor for a few milliseconds.

I think that the "massive lump of coal" construction that is typical of a carbon comp resistor is far more tolerant of this intaneous overload. The average power dissipated barely warms the resistor. The Metal oxide resistor doesn't seem to tolerant of this brief overload, neither does a metal film resistor. One started sparking around its surface and died in a few seconds. The amp exhibited a rather cool distortion (guitar amp) sound as the resistor was dying. This sound might be marketable to the right metal band! The second resistor just went poof with a small flash, a puff of smoke and a black stain on the PC board.
 
How many turns are we talking? If its ten turns or less...

For an audio power amp where the screen gain is low, they work just fine. Most of the resistors that I have dissected are 6 to 10 turns. So we have maybe 50 nH in series with at least 100 ohms. This could resonate with the screen capacitance of some tubes, but the loaded Q would be too low to support oscillation.

Some deliberately wrap that many turns on a stopper.

This is common practice on the plate of an 807 or a 6146. This creates a VHF choke with lousy Q to keep the tube from oscillating in the VHF region.
 
Member
Joined 2006
Paid Member
Has anyone tried the ceramic composition types like the OY series from Ohmite or the HPC series from KOA as an alternative to the 2W Carbon Comps for screen stopper applications?

Ohmite OX OY

KOA HPC

Construction appears to be carbon in a ceramic binder. No spiral cuts.

They're about a $1 USD a piece for the 2W size at Mouser. I plan to pick up a few to give them a try on my next Mouser order (which probably won't be for awhile).
 
Folks,

Let us just examine a few values here. I would respect Tubelab's guess of 50nH - he is seldom wrong. This would equal 100 ohms at .... 1 GHz. The inductance of several leads inside the amp might be far higher than of metal film resistors.

I have even used wire-wounds as screen stoppers - Q of about 1.1 - and never had any trouble with this. I did not measure inductance on metal oxide and the like; do not have the equipment. But I did exchange carbon with w.w. and had the merest twich of a peak on a square wave. This particular amplifier had a slight peak at 90 kHz before rolling off moderately fast. In the region 90 - 150kHz where the phase angle becomes large I had no visible difference between carbon and any metal type on the scope display. (This not only for screen resistors but also anode loads for pre-stages.)
 
Administrator
Joined 2004
Paid Member
Hi tubelab,
How would the resistor know the absolute voltage from the resistor to ground if it is suspended in free air, or mounted on a PC board with no nearby traces, which is the case for the Simple SE.
It wouldn't, and that is not what the voltage rating is for.

The voltage rating applies to total voltage across the resistor and assumes no other leakage path (imagine a double sided PCB with ground plain on top!). The failure mode is probably arcing between spirals. Your observations would seem to support that. Once the part tracks, it's done. The resistance would drop during an arc over and any tracking would permanently lower the value. So now you have a fraction of a resistor in circuit trying to support a higher current. This part now also has a much lower wattage rating. Complete failure is not a surprise.

Your other hypothesis seems likely. A solid composition type resistor would definitely handle surge currents better than any film I can think of. The current density would be far lower in the composition part. I've even seen very large flat wire wound 10 K resistors open up quickly when you have voltages above 550 VDC or so. Very impressive to witness. Try that in the dark!

Hi kenpeter,
Any inductance from higher value resistors that are wire wound, or helical cut, is not significant at all. I do measure the inductance of metal film resistors before using any new manufacturer. I did find at least one that had significant inductance. That shocked me, but also showed that you can not take anything for granted. The particular resistor I found had been installed by another technician in a quest to improve an amplifier at great cost to the customer. The original tech could not figure out why the amp broke into oscillation. He knew that metal film resistors were not inductive. Well, that one was.

In general, most resistors will have a Q that is far too low to cause any problems unless you are operating far above the audio band. For amplifiers that have gain around 1 MHz or higher, it's probably worth testing the part. Otherwise you are worrying about a very small detail.

Hi drj759,
They sound promising for certain applications. How noisy are they? Mind you, as screen dropping resistors, that probably matters not.

-Chris
 
Member
Joined 2006
Paid Member
anatech said:

They sound promising for certain applications. How noisy are they? Mind you, as screen dropping resistors, that probably matters not.



I'm curious about the noise performance too. I thought that with the solid body construction and absence of spiral cuts that they might have comparable surge performance as the 2W carbon comps. They’re at least half the price of 2W CCs (when you can find them). Hopefully my discretionary budget will recover soon and I’ll be able to get a few to try near term.

Edit: Looks like SpreadSpectrum had the same idea on the Ohmites (missed that comment earlier): Post #11
 
The voltage rating applies to total voltage across the resistor and assumes no other leakage path

The voltage rating also applies to the dielectric strength of the coating. The manufacturer has no way of knowing if the resistor is in direct contact with the chassis, another resistor, or an uncoated ground plane on a PC board (or worse a curious finger). Of course the rating applies to the terminal to terminal voltage too.

The failures that I witnessed were in a Simple SE PC board. I usually mount resistors 1/8 to 1/4 inch above the surface of the board, and there are no traces or ground plane under the resistors. I was using 150 ohm 1/2 watt resistor that was rated for 200 volts in series with the screen grid. There is no way I could get 200 volts across a 150 ohm resistor without some real fireworks! 1/2 watt is often called for for screen stopper use, and "normal" dissipation is below 1/2 watt.

I have run a bunch of different tubes through the Simple SE under many different conditions. The screen resistor, and everything else remained cool until I put in some JJ EL34's, and plugged my guitar preamp into the amp. Things were still OK, until I dialed up my favorite Jimi preset. After about a minute of this the first resistor started sparking out, some cool distortion ensued, and then silence.

I repeated the experiment. First, my old blow proof Chinese 6L6GC, no problem, dive bombing feedback, toggling the pickups on and off on a Les Paul, nothing bad happened. Ditto the KT88's. Enter the EL34. As soon as I held the guitar in front of the speaker and yanked on the pickup switch, the screen resistor acted like a micro Chernobyl. The JJ EL34's are not at fault, they are still hapilly alive 2 years later.

I put some 2 watt metal oxide resistors in that board, and they are still alive. This is my test board that has been shown on these pages torturing various tubes, including the incandescent 6BQ5's. I have blown the board up a few times, usually because I got carried away with my variable power supply and cranked 550 volts through some poor tube and blew the cathode resistor or bypass cap. The screen resistors have never fried since I put the metal oxide parts in. They are a bit discolored though.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.