Yes, but those you've been linking to aren't line input SMPS's.
Best regards!
what about something like this: LT1083CP 7A Tube Amp Filament Voltage Regulator Adjustable Power Supply Board 699973158209 | eBay
Yes I know. As I said, if you have a SMPS wallwart from an old phone that puts out 5V you can use a boost module to get up to 6.3V, if you have one above 6.3V then you use a buck module. Especially the phone ones can be found at thrift shops for a couple of bucks. They are roughly 1.0A. If you want a line powered one you are probably going to have to pay more than a buck for them.
Ah. So you've taken an already high-gain design, and raised the gain quite a bit higher with your modifications. This will certainly make it more challenging to make the amplifier stable, and also to manage hiss and buzz.
The question in my mind is whether your amplifier is actually unstable and oscillating continuously, or not. The radio reception problem suggests that it might be.
An oscilloscope would probably tell the tale immediately, if you have one. If not, if you have a radio try taking it near the amp, set it to MW, short-wave, and FM bands to see if you can pick up any squeals or weird noises from your guitar amp. Old-fashioned analogue TVs can be useful RF oscillator detectors too, though few of us have one around these days.
Typical input capacitance of a half-12AX7 gain stage is about 100 pF. A 34k resistor with that will only roll off treble above 50 kHz or so. So I think the grid-stoppers are blameless.
However: feed the whole thing from a 1 meg volume pot (like your preamp volume pot), and that can really change things. The pot itself can have up to 250k source impedance, and combined with the 100 pF input capacitance, that can roll off frequencies above about 6 kHz. You can certainly hear that, even with a guitar speaker...though many guitarists are actually happy to have these harsh-sounding upper frequencies removed!
That treble loss is probably why Marshall put those "bright caps" in the circuit - the ones you removed when you discarded the attenuators.
I would suggest putting the grid stoppers back, to help with the stability problems, and maybe trying a lower-value preamp volume pot to solve the treble loss problem. Even a 500k pot (instead of 1 meg) should do the trick - now only frequencies above 12 kHz will be rolled off, and there is virtually nothing above 12 khz from a guitar speaker.
Nothing obvious leaps at me, other than the fact that you have a trunkload of gain in that amp, and the chosen construction method places dozens of components close and parallel to each other, and then dangles several inches of closely-spaced wires off the ends of many of them, all of which increases stray capacitance and stray coupling between components.
I know this is one of those "classic" construction methods, and much praised on many guitar amp forums and websites. But it creates so much potential for electronic instability that I'm a bit amazed it ever works, even in low-gain Fender and similar designs.
-Gnobuddy
I am not a big fan of dragging the DC GND thru the chassis and connecting to multiple locations. I took an SS approach and I star grounded everything, and from that point only, I ran the DC GND to chassis and to AC GND. Maybe it was common approach back in 60's using chassis for GND. I see from the picture you posted, you used the chassis for DC GND a lot. The less current goes thru the metal enclosure the better, I think. I use my chassis only as a shield.
I agree with cobretti. The photos indicate multiple connections to chassis - the schematic doesn't show what is happening to the signal ground points - sort of back to post #10 suggestion.
JimvDB, if your aim is still a high gain amplifier then I'd suggest reading up on distributed star grounding and rewiring the amp grounds, and then confirming stage noise/hum levels by grounding each grid in turn and using at least a scope, and preferably a spectrum analyser. Then you might be in a position to go back and look at heater powering and how to do that adequately for your application (either with AC heaters and heater elevation and tuned humdinger, or some form of DC heater).
JimvDB, if your aim is still a high gain amplifier then I'd suggest reading up on distributed star grounding and rewiring the amp grounds, and then confirming stage noise/hum levels by grounding each grid in turn and using at least a scope, and preferably a spectrum analyser. Then you might be in a position to go back and look at heater powering and how to do that adequately for your application (either with AC heaters and heater elevation and tuned humdinger, or some form of DC heater).
Same here.
Among other things, this means using an input jack that's electrically isolated from the chassis - either a plastic jack, or one mounted with an insulating shoulder-washer so that input jack ground is not in contact with chassis ground.
My goal is to end up with no audible hum when the input jack is shorted - only hiss (and as little of that as possible, but that has nothing to do with grounding.)
-Gnobuddy
I made a similar suggestion in post #16.
I have a suspicion that it's going to take more than grid stoppers to solve this particular amp's problems, though. It's beginning to look as though the grounding scheme isn't ideal.
-Gnobuddy
Thanks you all for your great replies.
I always get radio reception with any amp because I live close to a radio transmitter. 5k6 grid stoppers did solve the radio reception for now. The amp sounds great but this hum issue bugs me.
So I tried to star ground this amp. I chose a point near the filter caps. Everything went great untill I got to the input jack ground and the first tube's cathode resistors/bypass caps.
Star grounding these gives more hum then the random grounding scheme I used before. This interesting result gives me a gut feeling the answer to the hum could be in these grounds.
What else can I try to ground these better? How can star grounding give more hum then random grounding?
I always get radio reception with any amp because I live close to a radio transmitter. 5k6 grid stoppers did solve the radio reception for now. The amp sounds great but this hum issue bugs me.
So I tried to star ground this amp. I chose a point near the filter caps. Everything went great untill I got to the input jack ground and the first tube's cathode resistors/bypass caps.
Star grounding these gives more hum then the random grounding scheme I used before. This interesting result gives me a gut feeling the answer to the hum could be in these grounds.
What else can I try to ground these better? How can star grounding give more hum then random grounding?
Excellent!
Is your input jack plastic, or otherwise electrically isolated from the metal chassis?
If not, the input jack will already be grounded to the metal panel it's mounted to, and when you add your star-ground wire, you'll actually create a big ugly earth-loop, which will cause hum and buzz.
Isolate that input jack from chassis, if it isn't already isolated...
-Gnobuddy
Excellent!
Is your input jack plastic, or otherwise electrically isolated from the metal chassis?
If not, the input jack will already be grounded to the metal panel it's mounted to, and when you add your star-ground wire, you'll actually create a big ugly earth-loop, which will cause hum and buzz.
Isolate that input jack from chassis, if it isn't already isolated...
-Gnobuddy
I was wondering along with the point. If you do have a metal jack that is grounded at the front panel, I have seen amps which had the first stage cathode resistor ground taken back to the input jack. I can not remember if the ground of the filter cap went back though.
The input jack is a plastic Cliff jack. It is so weird that the input jack and cathodes don't want to be grounded to a good 0v reference.
Is there any trick that lets me isolate them from the star joint yet still be connected? What I am thinking is that the chassis adds resistance to the connection. Would it be silly to add lets say a 10 ohm resistor?
Is there any trick that lets me isolate them from the star joint yet still be connected? What I am thinking is that the chassis adds resistance to the connection. Would it be silly to add lets say a 10 ohm resistor?
Oh, you certainly want the input jack and cathodes grounded to a good 0 V reference!
Factor #1: when you have multiple ground connections at different physical locations to the same chassis, small currents will flow through the chassis itself. Those currents flowing through the small resistance of the chassis metal will generate small voltage differences at all those points. So all points on the chassis are NOT good 0 V references now - instead, they have small noise and hum signals on them.
Factor #2: An amplifier can't tell the difference between an input signal at the input grid, and the same input signal at the input cathode - the triode amplifies whatever signal appears between its cathode and grid, and doesn't know which, if either, is zero volts.
Factor #3: You've built a high-gain amplifier that responds to very small voltages; low-gain vintage Fender amps were designed so 20 mV of input signal would drive the amplifier to full output power. Your amplifier will probably only need about one millivolt of input signal to produce (deafeningly loud) full output. Even as little as ten microvolts of noise would be only 40 dB lower than that - and 40 dB is a very poor signal to noise ratio.
Put those three factors together, and you see the problem. Poor grounding layout can easily cause a few microvolts of hum or buzz signal at some points in the chassis metal, and that's all it would take to cause your high-gain amp to hum and buzz.
Part of the cure is to make sure there is only one single point of your circuit that's connected to the metal chassis. This ensures that there are no noise currents flowing in the chassis.
Another part of the cure is grounding all the various parts of your amplifier together in such a way that the ground wires themselves don't cause the same problem that the chassis was causing - you don't want large noise currents flowing through any ground wires that connect to the sensitive input stages of the amp, generating noise voltages in the ground wiring itself.
That's the basic idea behind the star-ground: everything grounds at one point, so the ground wires from the input stages don't carry big noise currents; the big noise currents from the output stage do cause voltage drop, but only in the ground wires to the output stages, and the output stages by themselves have little voltage gain.
Star grounding sounds great in principle, but in practice, it's impractical except for very simple circuits. The circuit you built is not that simple - you must have at least a dozen different points in your circuit that all need to connect to ground in your amp. It's not practical to cluster twelve or fifteen ground wires all together at one point, so true star grounding is probably not going to be an option.
Instead, for a more complicated circuit like the one you built, you need a different solution. One that works well is a "star of stars", where, for example, you have a local star ground that connects the input jack ground, the input triode's cathode resistor, the input triodes cathode bypass cap, and the input triode's power supply filter cap, if any. Now you run a single wire from that small local star to the main star ground.
Merlin Blencowe has very kindly made his excellent chapter on grounding available on his website, at no charge. He explains the "star of stars" concept, as well as a lot more: The Valve Wizard
There are situations where a small resistor in a signal ground can help. But it is also absolutely vital that you never, under any circumstances, add any resistance to the safety ground, the wire that runs dangerous fault currents back to the ground pin of your AC outlet.
Adding resistance in that path removes the safety advantage of 3-wire AC, and is not only dangerous, it's almost certainly illegal as well.
Imagine, for instance, that you added a 1k resistor between the input triode's cathode and the safety ground. Now let's suppose that triode suffers an internal mechanical failure, perhaps as a result of being jarred or overheated - its anode short-circuits to its control grid, just a few thousands of an inch away inside the glass envelope.
Now your guitar cable is directly connected to the dangerously high anode voltage of that triode, and the dangerous fault current that would normally flow harmlessly to ground is blocked by the 1k resistor; that current will now flow to ground through the guitarist's body instead.
Anode-grid shorts do happen, though not frequently. But other circuit failures could cause similar problems; for instance, an accidental short caused by a stray blob of solder or overlooked whisker of wire between B+ and the control grid of the input triode. If the safety ground is in place, the ground wire will save the guitarist's life. If dangerous resistance has been added to that ground wire, the guitarist is going to be shocked, and maybe injured.
So is ten ohms of added resistance safe? I don't know (it could be dangerous if enough fault current manages to flow through it.) I certainly don't want to be the human guinea-pig that finds out the hard way. Betcha you don't either!
-Gnobuddy
I have completely wired the grounds again to a different point on the chassis. All grounds go go one point on the chassis.
Result is more hum than random grounding. This goes against everything I have read about the subject.
Can I conclude my grounding was good in the first place or is there anything left to try?
Result is more hum than random grounding. This goes against everything I have read about the subject.
Can I conclude my grounding was good in the first place or is there anything left to try?
Frustrating, but remember, hum problems are one of the harder ones to fix, because there are many possible causes:
- Hum can be caused by power supply ripple.
- Hum can be caused by the lack of electrostatic shielding (unbroken grounded metal all around the amp.)
- Hum can be caused because the power transformer is too near an input valve, and wiring is picking up the radiated magnetic field.
- Hum can be caused because the PT is too close to the OT, and the magnetic field from the PT is coupling hum straight into the OT. (My Blues Junior had this problem, and would hum with *all* valves removed!)
- Hum can be caused by AC mains wiring that is located too close to sensitive parts of the circuitry.
- Hum can be caused by AC heater wiring.
- Hum can be caused because of poor grounding layout.
For example, if you move your hand near the input valve (without touching it), does the amp hum more? If so, you may need one of those grounded metal slip-on shields on V1.
What happens when you pull the power plug out of the wall outlet? Does the hum stop during the initial second or two that the amp is still live (running off the fully charged power supply caps), or does it continue to hum until the bitter end?
If it hums the whole time, your hum problem is not caused by the incoming AC power. If the hum stops immediately when you unplug the mains power, you're most likely dealing with one or more of issues 1,3,4,5,6.
You already did a test with your 6V lantern battery showing that the hum wasn't caused by your AC heater power to V1, so you can rule that out (though I suggest repeating that test now that you've changed ground wiring.)
Does your amp hum with output valves removed? If so, you're dealing with magnetic coupling straight from PT to OT, and one of them needs to be moved, or re-oriented.
And so on; you have to invent tests that eliminate possible sources of hum, one at a time, until you find the culprit.
Have you read through Merlin Blencowe's chapter on grounding? I found it very helpful with a humming DIY amp build of my own.
-Gnobuddy
I grounded my JCM800 like this, see the picture. The star ground is kind of exaggerated. It is of course impractical to bring all individual component to one GND point. But if all components are laid out and wired in straight line as seen on the picture I highligted red, there shouldn't be a conflict with current flow. The last point (or first, depends how you look at it) is connected to chassis, and that point is also connected to AC GND in my amp, and, I also have a center tap from heater windings connected to that point, and there are no hum issues. I did not solder any wire to back of potentiometers either. Potentiometers are inherently grounded to chassis thru locknuts and mechanical securing. If you do start soldering componets on them, the current, once again, will stop flowing "in line" fashion and will be going thru the chassis. I have a feeling that in this high gain amplifier, the wires are picking up the magnetic field from the power transformer, penetrating the enclosure. As Gnobuddy mentioned earlier, a few hundreds microvolts or even less, I would say few tens of microvolts will definitelly come all the way thru to the last stage and be amplified. And I am pretty sure the magnetic field of the power transformer will do that. Sometimes the magnetic field travels thru the chassis as well. Depends on orientation and other things. I measured as low as 2mV 60Hz on my 8 ohm speaker and that was very audible! Convert to standard proven design and you will be headache free.
Attachments
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.