I'm working on a Kolster K-20, a 7-tube, AC-powered set, c. 1928. This is an interesting radio with several nice bells and whistles. Schematic: http://www.nostalgiaair.org/PagesByModel/092/M0040092.pdf

I got it sort of working without much effort, but it performed very flakey - erratic sensitivity, oscillation, etc. This set has three #26 RF stages, each fed through a grid-stopper which, in turn, is bypassed by a small cap. Lo and behold, I found that all three grid stoppers were wide open.

These were wire resistors with fine magnet wire wound on a rectangular cardboard form. These type resistors are usually OK, and they carry hardly no current - so they couldn't have "burned" out. I figure they all got corroded, perhaps at a connection. I haven't disected them to investigate how they failed.

The schematic doesn't show the value of these resistors, so I "guessed" and shunted the open stoppers with 680-ohm resistors. The set is behaving much better!

I took a look at some other sets that use the looser resistors in the RF section. The Radiola 16 uses 201a tubes and 600 ohm resistors. The Radiola 17 uses #26 tubes and 1000 ohm resistors. Your guess is most likely close enough. From what I remember they put those resistors in there to prevent oscilation. If it has a tendency to oscilate increase the value.

MRO

Thanks, Mark. I'll let it play for a while, and if there is any significant oscillation, I'll try jacking up the grid stoppers to 1K ohms.

In this case, with the stoppers wide open, there was no direct path for the RF signal to get from stage to stage, except through the puny caps that bypassed the stoppers. So, in essence, the grid stoppers were infinite resistance, and it oscillated. I guess with the grids flapping around with a high impedance input, weird things should be expected. It's a wonder the set played at all.

What this experience shows is how much a radio WANTS to play, even if it is seemingly crippled beyond belief.

Doug, the resistors and caps make up a filter to suppress the higher frequencies in order to prevent oscillations due to excessive feedback. They are not supposed to do anything at the lower frequencies, which are less prone to oscillation, and need all the gain they can get. Therefore, if you experiment to find the correct resistance, do your testing at the highest signal that your set will tune.

Resistors are not usually wound with magnet wire, but some of them look like it if enamel was used for insulation. However, many of them were space wound further apart and did not have insulation. The resistance wire used on them cannot be soldered, so the wire was often attached with an eyelet crimped to a piece of cardboard. I have found a few of these with bad connections at the crimp, but usually the wire is broken. Some of the resistance wire made back then was not very reliable and began breaking with age, sometimes in many places on a single resistor. I have seen numerous potentiometers that had so many breaks that they stuck out like a porcupine. Even a few rheostats with heavy wire were prone to breaking. If you are fortunate and one of your resistors only has a single break, just measure the resistance of the two pieces to determine the correct value for a replacement.

A radio such as yours will still work to some degree without the resistors, but as you say, very erratic. Without a resistor, excess electrons build up on the grid and a tube becomes blocked. When enough electrons bleed off, the tube goes back to work. When the tube goes back to work, even the puny caps in your set are more than enough to pass the entire RF signal without a resistor to suppress it, thus causing intermittent oscillations....br

Thanks, Billy. I don't have the set back in its cabinet yet, so I guess I'll disect one of the original stoppers and see if I can measure the resistance.

I removed the paper wrapping from the stoppers and inspected them more carefully, insitu. The resistance wire isn't quite like a porcupine, but it's broken here and there. As best I could determine, the resistance is around 800 ohms (compared to the 680-ohm resistors that I had replaced them with).

I decided to try 1K stoppers (the next size resistors that I had available). I do believe that the set is a little better behaved, particularly toward the higher freqs.

I'll let the set play for a while on the bench (listening to the Bears game), and then button it up.

Hello all. What are grid stoppers? Are these grid leak resistors?

Thomas

Thomas: No, the grid stoppers are resistors in the input line to a grid. Typically used on RF stages of TRFs with transformer coupling. They usually are several hundred ohms. They settle down oscillations. There's a little black magic involved and I don't think their size is determined mathmatically, just by trial and error?

Thomas know this, but grid leak resistors are a different animal. They are used in a triode that is used as an audio detector. The grid of the triode is really unneeded - all you really want is a diode. But, you have to hook up the grid, through a resistor, to ground to keep the grid from building up a static charge of electrons, and cutting off the tube. Grid leaks are usually several megohms.

More: I think grid stoppers were used mainly in TRFs employing simple triodes, such as the '01A or 26, in the RF sections. These tubes had a relatively large internal capacitance between grid and plate, which could lead to oscillation. Grid stoppers helped reduce the oscillation. The Neutrodyne circuit helped too, but that required a license from Prof. Hazeltine.

Later, around 1929-30 or so, manufacturers began using screen grid tubes. The screen, positioned between the grid and plate, reduced the tube's internal capacitance. At that point, I think grid stoppers and the Neutrodyne circuit became obsolete.

Doug,
This was a way around patents. A lot of manufacturers paid royalties to use the neutrodyne patent which used feedback capacitors to equalize grid to plate capacitance of the tubes. Look at the schematic for a Freed Eisman NR-5 and you will see it. Companies like Stromberg Carlson in their Model 1-A also used it. Atwater Kent absolutely refused to pay royalties and that's why sets like the Model 40 series with 26 triode tubes use the grid resistors. They used 600 ohms but they were not that critical. Sparton went a different route refusing to pay royalties. They used bandpass tuning where three or four tuning circuits were used ahead of five untuned RF triode stages. All these were schemes to avoid patent royalties. As mentioned, screen grid tubes eliminated most of the problems within a year or two.

Fred R

:More: I think grid stoppers were used mainly in TRFs employing simple triodes, such as the '01A or 26, in the RF sections. These tubes had a relatively large internal capacitance between grid and plate, which could lead to oscillation. Grid stoppers helped reduce the oscillation. The Neutrodyne circuit helped too, but that required a license from Prof. Hazeltine.
:
:Later, around 1929-30 or so, manufacturers began using screen grid tubes. The screen, positioned between the grid and plate, reduced the tube's internal capacitance. At that point, I think grid stoppers and the Neutrodyne circuit became obsolete.

I see what you're saying. The grid stopper resistor is kind of like a grid leak resistor in how it's wired (across the condenser), but it's a lower value? I think this is what I'm seeing in the schematic. I saw a few grid condensers in the RF section with resistors across them. I couldn't find values listed, but assume that these are lower value resistors instead of 1 megohm resistors (or thereabouts).

Interesting. I wish I had money so I could build things like this or maybe own more radios....or maybe more books. Still, with a site like this I learn things all the time anyway.

Thomas

I never heard that terminology for this resistor either Thomas, so don’t feel like the Lone Ranger. In my neck of the woods, we call the resistor a grid suppressor. I suppose you could say that the resistors help to stop the oscillation, but you could also say they suppress the gain to reduce it. The remainder of this post is not necessarily directed to you though Thomas, and also applies to some of the other comments that have been made on this subject.

I believe the AK 20 was the first manufacturer to use these resistors, but some of the construction articles in various publications also used them with home built radios. Although the resistors reduced feedback (oscillation) at the higher frequencies where it was needed, they were not quite enough, and other methods were used in combination with them. Atwater Kent’s answer for this was to install the tuning coils on the back of the tuning condensers. This reduced the gain on all frequencies by cutting the magnetic fields, even on the lower frequencies where it was not needed. The Neutrodynes also sacrificed the lower frequencies, so they did not have an advantage in this respect

Installing the coils in this location worked so well that a lot of other manufacturers used this method without resistors, but some of them got ridiculous with it. They went to great lengths to design what they thought of as highly efficient, low loss coils, then destroyed it all by cramming them on the back of a tuning condenser. The early Freshman Masterpiece is a good example of this mentality.

A construction article in one of my old radio magazines describes the action of the grid resistor fairly well, but I don’t remember exactly what it said. Basically, it stated that a resistor of the correct value does not have time to respond at the lowest frequency of the broadcast band. In other words, it acts like a straight piece of wire. As the frequency increases, it becomes more and more effective, thus automatically reducing the gain up to the highest frequency of the broadcast band. More resistance is needed to completely control feedback at that frequency, but if it is increased, it will reduce the gain at the lowest frequency. Or something like that according to the way I understand it. Not mentioned in the article is that a certain amount of capacity exists in the tuned circuit, so it has to come into play, at least to some extent.

This brings us to the condensers in Doug’s Kolster. I do not have any experience with this method, so probably should not analyze them on this forum. However, according to my way if thinking, they seem like a good idea to me. An arrangement such as this has to be tuned to some specific frequency, so I know what I would try if designing such a thing. I would calculate the values of the condensers and resistors to resonate at 550kc or thereabouts. The resistance should be almost zero at this point, or about the same thing as a straight piece of wire. Any stations tuned away from resonance should increase the resistance so that the circuit becomes a filter that automatically rejects the signal more and more as the frequency becomes higher. Sounds to me like this would be better than a plain old resistor, but who am I to say?

Of course, the screen grid tube eliminated these gimmicks along with others that have not been mentioned here, but it did not eliminate feedback. It just made it easier to control. However, some of the older RF transformers could probably be used with these tubes without doing anything to prevent feedback, because they were designed for use with tubes having more capacity between the elements. Since the screen grid tubes allowed more gain without feedback, most manufacturers took advantage of their ability, and designed coils that put them almost right back where they were with the older tubes. All they had to do is add a few more turns on the primary to come closer to the plate impedance, thus allowing the tube to amplify more. This extra gain was enough that most of them did not worry much about slight losses at the lower frequencies, nor were they very concerned about slight losses due to alignment of ganged tuning condensers. To control feedback, most of them took advantage of another feature of the screen grid tube. They simply used the voltage on the screen to control it. Some radios went one step further and had a potentiometer to one of the screens. This pot allowed a gentle oscillation that was very easy to control at frequencies that were prone to do so, and full amplification on all the other frequencies.

The grid leak and condenser on the detector has the same arrangement, but this tube is a different hoss. Not only does it rectify the signal, but it also amplifies it at the same time. If it was replaced with a diode, then an additional audio stage would be required. However, if the grid leak is open, the same thing happens on this tube that does on an RF tube with an open grid resistor. Either tube can become blocked due to excess electrons on the grid.

This was not always the case with a few tuners back in the wireless days, though. They did not have a grid leak, and worked fine without one. It is believed that the grid condenser had enough leakage to compensate for it, but this has not been proven by anyone that I know of. Some of the sets back then used a pencil mark for a grid leak. This mark was between two contacts on the front panel that were protected with a small removable cover. If the operator needed less resistance, he made the mark heavier or wider. If less was needed, he erased some of it. Crude, but effective.

On some of the battery sets that I design and build, I may eliminate the grid leak and condenser if it gives better performance without them. This is done with a “C” battery, or in most cases with a single cell. I’ve never made a study of why this can often work better, but since the electrons through a cell only travel one way, they do almost the same thing as a rectifier. Maybe that has something to do with it, I don’t know....br

Billy, your theory about sizing the grid stoppers/supressors is interesting. I didn't think to measure the value of the caps that are in parallel with the stoppers, and now I've got the whole she-bang back together - so I'm unable to test your theory. In any case, the set is playing satisfactorily.

All of the grid stoppers I've encountered are wire wound, some on a flat form, some on a small cylindrical form. (Back in those days, I think most if not all low-value resistors were wire wound.) I read somebody's theory that the small amount of inductance in the original resistor plays a role. But, I've always replaced them with ordinary film or composition resistors.

Regarding grid leaks, I've encountered TRF sets that played OK with the grid leak resistor removed or open. Other times, I've gotten a putt-putt sound out the speaker, which I suppose is the detector grid charging up until its voltage is high enough to leak off through some path.

Doug, all resistors, whether wire wound or not, tend to supress higher frequencies much like a choke, though not to the extreme of a choke. I can see a wire wound resistor having more AC impedance, but you are right about new resistors working in the old circuit.

Interesting what you say, Bill. I have never heard the grid stopper terminology before. Speaking of unwanted oscillation, though, a long time ago I repaired a 4 tube AC TRF for a neighbor of mine. It was one of those Century models with two 24 RF tubes, a 47 output, and an 80 rectifier. At the time I assumed that the thing used regeneration, because when the volume control was turned up, oscillation occurred. None of the tubes were shielded, though the RF coils were. Looking back at the circuit, though, it is obvious to see that the circuit employed no regeneration, and the oscillation must have been the unwanted type which you talk about. To a degree it did work like regeneration, though, because signal strength was greatest just before the set broke into oscillation.
Thomas

You were probably right the first time about the Century. A radio with that tube line up would almost have to use regeneration, even though the circuit may not show it. There are a lot of different ways to control regeneration, and from your description, I’d say the volume control did just that. All they would have to do is design the RF amplifier to oscillate at all frequencies, then use a potentiometer somewhere in the circuit to knock it back. Probably somewhere in the first stage....br

Well, here's the schematic: http://www.nostalgiaair.org/PagesByModel/645/M0002645.pdf

It's a trick to read. I can't see any regeneration. The set wasn't much of a performer. All it did with an indoor antenna was pick up local stations, and not extremely well, either. The tube line-up is 24, 24, 47, and 80.

Thomas

I looked at the schematic. You’re right. It’s drawn rather crude, which I suppose is fitting for a radio with a circuit such as the Century. However, it is very much what I had imagined, except the potentiometer controls the gain on both RF tubes instead of one, as I had guessed.

I only glanced at the schematic, so this is the way I saw the circuit without study. Any TRF amplifier with 2 RF stages will oscillate with decent RF transformers. This excessive feedback can usually be eliminated by lowering the screen voltage. The Century probably has a screen voltage that almost eliminates it, but not quite. Therefore, it needs a wee bit more of something to completely eliminate it. Using a pot to control the cathodes of both tubes is a slow gentle way of controlling it, so this method should work pretty good. In other words, it makes a good regenerative control. Many of the radios during this time period used a pot on the antenna as a volume control, and the one in the Century serves double duty and does this too. Not a bad idea in my opinion.

All early AC powered radios had a problem controlling the volume. They needed to amplify the weakest signal on the antenna and also a strong local signal, which could cause excess distortion or sometimes even worse; excess feedback in the form of oscillation. A pot can easily control this, but it left the audio amplifier running wide open. A good strong audio amplifier needs a pot to control the input in order to reduce hum, noise and distortion, so most of these early radios really needed two volume controls. The automatic volume control solved this problem because it controls both the RF amplifier and the audio amplifier. The Century doesn’t have an audio amplifier worth a poot, so this is not much of an issue.

Actually, this radio would work pretty good with a long wire antenna, which was probably used with most of these sets back then. A single tube regenerative tuner has enough selectivity and gain to receive distant or weak signals on a good antenna, and the Century can do the same thing since it is a regenerative set.

It works alright. If you string an outdoor antenna it does well. No matter how you tune the trimmers, though, it lacks uniformity over the dial. It's an alright set. The real problem was the rather interesting speaker. The cone is made of rubberized cloth. Well it is stiff. I used to soak the edge in acetone, which would enhance the bass really nicely, but this would harden in a few minutes. This was about a decade ago, though, so perhaps to-day I could have washed all the rubber out of the cloth around the edges with acetone. This would allow re-rubberizing the cloth. Hmm....not worth the trouble. Some other time.

I discovered something recently that's kind of nice to know. When you have a stiff speaker cone that lacks bass, you can rub olive oil into the corrugated outer edge of the cone. This softens the paper well here (it also makes it less prone to tearing). It's almost as good as wetting the outer edge, but not quite. When you wet the outer edge, though, the effect only lasts for about a half hour. The olive oil trick may last much longer (a couple of years maybe). It is not wise to put olive oil on the main part of the coil, as this will cause it to soften, and it may not move so much as a unit after this treatment, which would mute some treble (or cause odd harmonics). You can try oiling the main cone if you wish. I've tried it on some spare speakers, and didn't like the results. Petroleum oils, by the way, will not soften the paper. Olive oil is the only one that works for me thus far.

Another trick I've tried, and don't know how long it'll last, is to first wet the outer edge of the cone and then rub in the olive oil. The oil sort-of retains the moisture, and the cone vibrates more readily. I am trying to think of things which are soluble with paper glue, that will retain some moisture for a long time, so as to keep the glue (and the paper fibres) flexible. I wonder how olive oil and dish detergent would work when mixed together. The detergent may make the oil more compatible with the water based paper products. It would be neat if someone sold a paper restorer.

Thomas

:It works alright. If you string an outdoor antenna it does well. No matter how you tune the trimmers, though, it lacks uniformity over the dial. It's an alright set. The real problem was the rather interesting speaker. The cone is made of rubberized cloth. Well it is stiff. I used to soak the edge in acetone, which would enhance the bass really nicely, but this would harden in a few minutes. This was about a decade ago, though, so perhaps to-day I could have washed all the rubber out of the cloth around the edges with acetone. This would allow re-rubberizing the cloth. Hmm....not worth the trouble. Some other time.
:
:I discovered something recently that's kind of nice to know. When you have a stiff speaker cone that lacks bass, you can rub olive oil into the corrugated outer edge of the cone. This softens the paper well here (it also makes it less prone to tearing). It's almost as good as wetting the outer edge, but not quite. When you wet the outer edge, though, the effect only lasts for about a half hour. The olive oil trick may last much longer (a couple of years maybe). It is not wise to put olive oil on the main part of the coil, as this will cause it to soften, and it may not move so much as a unit after this treatment, which would mute some treble (or cause odd harmonics). You can try oiling the main cone if you wish. I've tried it on some spare speakers, and didn't like the results. Petroleum oils, by the way, will not soften the paper. Olive oil is the only one that works for me thus far.
:
:Another trick I've tried, and don't know how long it'll last, is to first wet the outer edge of the cone and then rub in the olive oil. The oil sort-of retains the moisture, and the cone vibrates more readily. I am trying to think of things which are soluble with paper glue, that will retain some moisture for a long time, so as to keep the glue (and the paper fibres) flexible. I wonder how olive oil and dish detergent would work when mixed together. The detergent may make the oil more compatible with the water based paper products. It would be neat if someone sold a paper restorer.
:
:Thomas

Hi Tom; Try olive oil and white vinegar. Vinegar softens old glues. Dan

Early resistors were wire wound for a few different reasons. The main reason is because that was about the only way they knew how to make a reliable resistor to carry current. They were also more accurate because all it took was a certain length of wire to keep them within manufacturing tolerance. At that time, all other resistors were a hit or miss concoction that may be affected by humidity, age, or whatever, and wire wounds did not have this problem. However, the main fault of a wire wound resistor back then was that it was limited to a fairly low value simply because they did not have a reliable resistance wire with a small enough diameter.

It’s true about the resistors back then having a small amount of inductance. They were wound on a relatively large form, which was about the same as a small tuning coil. They knew this way back when, and that is the main reason for the flat wound type, which have no inductance. If you can picture this in your mind, all the wire you see on one side of this type resistor is running at a slight angle. These same angles continue on the opposite side, and if the resistor was transparent, you would see that they are all crisscrossing each other. If the wire form is thin enough, the fields created will null each other, much the same way filament wires do when twisted together.

So flat wound grid resistors were better back then because they were accurate, didn’t change in value, and were non inductive. Modern film resistors do the same thing or better, so they make ideal replacements. As a matter of fact, the low wattage ones are so small that they can easily be hidden on the back side of most flat wound resistors, using the same connections. If the original was covered with paper, then so much the better. Replacing or reusing the original paper if possible would make them almost invisible.

Yeah, some detectors work without a grid leak, some don’t and some make weird sounds. Some just get louder when the electrons bleed off, and others have what I think of as a dripping sound, which I think is appropriate for the word grid leak.

My intentions were to respond to your post sooner Fred, but I haven’t been feeling up to it until now. Most of what you said about the neutrodyne patent holds true, but Sparton was getting around this patent in 1926 with their battery powered models 5-16 and 5-26. Later on in this year, Sparton began selling the AC5, which was an electrified version of the same circuit that used tubes manufactured under the McCullough patent.

The unusual circuit that Sparton came up with avoided the Alexanderson patent concerning cascading TRF circuits, which RCA was holding. When RCA began suing all TRF manufacturers, including the Neutrodynes in 1927 for infringing on this patent, Sparton’s pre-tuner or passive TRF circuit eliminated the cascading amplifier. Atwater Kent developed the Model 50, which also circumvented the Alexanderson patent in about the same way as the Sparton.

Although the Neutrodyne was becoming obsolete around this time because of the screen grid tube, you will see Hazeltine tags on some of the later radios, mostly because of the automatic volume control patent. The AVC was invented by Harold Wheeler in 1925 while he was working for Alan Hazeltine, so the Hazeltine Corp owned the rights. I would like to build a radio that used his early AVC circuit, so if anyone has any information on it, or better yet a schematic, I sure would appreciate you letting me know....br