Greetings Friends. I've been tinkering with the Shure M65 Phono Preamp DIY Kit from ZeroZone and have found it to be an excellent performer, especially for the price. Depending on enclosure and tube selection, I can build one for $200, give or take. Here it is in a Hammond 1458 VE3.
I like these boxes because they're well ventilated and easy to work on (the end plates are aluminum) but maybe that good ventilation is letting in too much RF.
As you can see see the PT is close to the board and the 12AX7 tubes are in plain sockets. I have replaced the plain output wire with shielded cable since this pic was taken.
At my home this preamp sounds great and has low/non-exist noise. However, at my buddy's place:
He's got cell phone antennas almost directly across the street from his house and this noise is present in his guitar amps as well, but it's much worse with this preamp. I assume that's because of the high gain in the circuit. A new Schitt phono pre is virtually silent in the same system.
So what should I try? Would shielded tube sockets help? I'm not sure they'd fit the box. So far we've tried wrapping the box in copper tape, EMF filter on the power input, nothing. Any ideas?
Thanks!
I like these boxes because they're well ventilated and easy to work on (the end plates are aluminum) but maybe that good ventilation is letting in too much RF.
As you can see see the PT is close to the board and the 12AX7 tubes are in plain sockets. I have replaced the plain output wire with shielded cable since this pic was taken.
At my home this preamp sounds great and has low/non-exist noise. However, at my buddy's place:
He's got cell phone antennas almost directly across the street from his house and this noise is present in his guitar amps as well, but it's much worse with this preamp. I assume that's because of the high gain in the circuit. A new Schitt phono pre is virtually silent in the same system.
So what should I try? Would shielded tube sockets help? I'm not sure they'd fit the box. So far we've tried wrapping the box in copper tape, EMF filter on the power input, nothing. Any ideas?
Thanks!
Depending on how slavishly they copied the original, you might start by adding grid stops to the first stage, and terminating the input cable for RF with a Zobel of series'd 75R and 220pF mounted right on the input jack. A build-out resistor of 1 or 2 K Ohms, mounted again right at the jack, wouldn't hurt. If you're lucky you won't have to try to filter the mains.
All good fortune,
Chris
All good fortune,
Chris
Thanks for taking a look. I neglected to include a schematic in my first post, here it is:
And there's plenty of space to put a 75R resistor and 220pF cap at the input jacks. Assuming a small MF resistor would do, what type of cap should I be looking for? And could you clarify Build-Out?
Thanks
Will
And there's plenty of space to put a 75R resistor and 220pF cap at the input jacks. Assuming a small MF resistor would do, what type of cap should I be looking for? And could you clarify Build-Out?
Thanks
Will
This schematic is full of ridiculous component values. B+ of 100VDC, R2 of 47K, C16 of 470pF? WTF?
A build-out resistor is an isolating resistor at the output, especially of an amplifier with loop feedback taken from the output, and goes downstream of the feedback takeoff point, so right at the output jacks. 220pF cap can be anything physically small; an NPO ceramic would be nice. The idea is to terminate the cables from the turntable to minimize standing waves.
All good fortune,
Chris
A build-out resistor is an isolating resistor at the output, especially of an amplifier with loop feedback taken from the output, and goes downstream of the feedback takeoff point, so right at the output jacks. 220pF cap can be anything physically small; an NPO ceramic would be nice. The idea is to terminate the cables from the turntable to minimize standing waves.
All good fortune,
Chris
Blame the original Sure schematic from 1960 , except that R2 47K unnecessary load , it is the same ...
The sourse of the noise should be investigated first , if it is present or not with shorted inputs and so on
The sourse of the noise should be investigated first , if it is present or not with shorted inputs and so on
No grid stopper on the input triodes, I would try that first. I've also put 1nF from input jack ground to chassis for RF, not sure how that's better or worse than Chris Hornbeck's zobel. In my phono preamp I also added a ground isolation circuit between circuit ground and house ground, and added a power line filter (maybe you already have one in that power input module.)
The picture looks like the input grounds are directly connected to chassis, but there's a ground loop introduced from output jacks to input jacks. Best practice is to remove this connection, and bring outputs hot and ground, from the board to the jacks with no other connection. The rule is: one and only one ground path.
All good fortune,
Chris
All good fortune,
Chris
Like Chris Hornbeck said but besides the already existing ground loop also not helping to fight HF/RF issues there now also is a direct connection between Audio GND and PE. It should be clear that the isolated RCA connectors don't make use of that isolation 🙂 Well there are spent tens of thousands of words on that. Only the casing should have a 0 Ohm connection to PE. Back to the basics of electrical wiring! This device needs RC filters at the inputs as usual today as almost all homes are full of HF/RF pollution and this certainly when RF transmitters are nearby. I hope RC filters are enough in this case and you could add an IEC inlet combined with built in mains filter to avoid stray in of RF via the mains cabling. It can very well be that the other devices in the chain also need that!
Philosophy: apparently all tube devices are (predefined) excellent performers despite the ever returning design errors/omissions that only seem to become more since the far east produces stuff for the DIY tube market. Bug hunting and finding errors is therefor essentially unnecessary isn't it? It would help to see bugs and error as bugs and errors and strive for best possible performance.
Philosophy: apparently all tube devices are (predefined) excellent performers despite the ever returning design errors/omissions that only seem to become more since the far east produces stuff for the DIY tube market. Bug hunting and finding errors is therefor essentially unnecessary isn't it? It would help to see bugs and error as bugs and errors and strive for best possible performance.
Last edited:
We are not discussing mains hum here, but RF interference. The larger you make the loop from the output cinch connector grounds to the chassis, the more effectively it will couple in RF. For good RF shielding, use an enclosure that makes reliable contact, connect the grounds of the connectors straight to the enclosure (no insulated connectors and no star grounding!), use an IEC connector with built-in RF filter, add some RC filters at the inputs and outputs straight after the connector and hope for the best.
A less invasive thing you could try first are clip-on ferrites on all input and output cables, as close to the amplifier as possible.
If insulated input RCA connectors are used a ceramic disc cap from the RCAs GND to the GND/PE point of the casing usually also helps. First the basics though as ground wiring is done erroneously. Simplest is to cut the unisolated GND wire between the black wire and the input RCAs GND and then test. It would surprise me that the RF coupling is worst at the outputs.
Last edited:
It's rare for vacuum valve amplifiers to suffer (significantly) from RFI. I can't help but think that input grid stops and doubling the B+, to get the input stage properly biased, would solve the issue. The circuit is crippled unnecessarily, making it more vulnerable than it could and should be. But all of the recommended measures will help.
All good fortune,
Chris
All good fortune,
Chris
Used to be like that but today things really are different. It is also in various frequencies/bands and sometimes at pretty high levels too. Even normal household stuff may be a hidden transmitter like for instance LED lighting.
It is so common that countermeasures should be in the average audio design. One can not ignore it anymore and prevention is better than curing.
It is so common that countermeasures should be in the average audio design. One can not ignore it anymore and prevention is better than curing.
Last edited:
Given the high levels of RF it sounds like the box needs strict attention to RF leakage (ie proper RF gaskets etc), and all the wires going into it need ferrite toroids at the very least. Currently I suspect a lot of RF is getting into the box and spraying everywhere. At the frequencies used by mobile phones these days you have to use microwave screening techniques, good continuous Faraday cages without slots or holes more than a few mm across. The coax signal leads into the box will need to have RF-tight braiding too I suspect. Note for RF screening the inputs must ground to the case, otherwise you are routing RF straight into the input section.
Thank you all for the good information and discussion. There's obviously a lot I still have to learn about the construction of these devices. In this case I am limited to the parts supplied, so the b+ will stay at 100v. I've been under the impression that linking all the signal grounds and having only one point touching the PCB was better, so mine look like this:
With a single wire connecting the Input Ground to this ground buss. But I should cut the wire between the Input and Output lugs and run a wire from the Output Ground to that? makes sense.
What is PE ? Power Earth? I don't have a lot of formal education in wiring basics. I get the safety earth connections, but those R-C filters to Earth, I dunno.
So far strategies include
1 a Zobel network, 75R resistor and 220pF cap in series at the signal input
2 build-out resistor, 1K resistor in series at output jack
3 Grid Stopper at Pin 2 on both sockets. difficult because the socket is soldered to the PCB
4 Ferrite Toroid on input wires. Compatible with shielded cable?
Thanks!
will
With a single wire connecting the Input Ground to this ground buss. But I should cut the wire between the Input and Output lugs and run a wire from the Output Ground to that? makes sense.
What is PE ? Power Earth? I don't have a lot of formal education in wiring basics. I get the safety earth connections, but those R-C filters to Earth, I dunno.
So far strategies include
1 a Zobel network, 75R resistor and 220pF cap in series at the signal input
2 build-out resistor, 1K resistor in series at output jack
3 Grid Stopper at Pin 2 on both sockets. difficult because the socket is soldered to the PCB
4 Ferrite Toroid on input wires. Compatible with shielded cable?
Thanks!
will
PE = Protective Earth, what prevents you being electrocuted when there's a fault.
Your ground wiring from the input connectors to the bolt is radiating RF straight into your box - phone signal wavelengths are measured in cm so parts of it may even be resonant at the RF frequencies involved. To keep out RF you need to ground the input sockets direct to the metal enclosure. This isn't the ideal for audio, but its what's required to block RF.
Your ground wiring from the input connectors to the bolt is radiating RF straight into your box - phone signal wavelengths are measured in cm so parts of it may even be resonant at the RF frequencies involved. To keep out RF you need to ground the input sockets direct to the metal enclosure. This isn't the ideal for audio, but its what's required to block RF.
The silver bolt in that picture is the ground lug, I scrape the paint away with a counter-bore bit.
You're saying they should connect straight to the ground where they are?
You're saying they should connect straight to the ground where they are?
Before adding all sorts of components , the grounding in the picture is obviously not good , input and output should be separated going through their shielded cable to PCB and the chassis tied to star ground , at the power supply main capacitor minus terminal .
And remove R2 47K from output , it is loading the preamp too much and is not present in the original schematic . I think it was stated that the load for this circuit should be no less than 47K and somebody not very smart understood that 🙂
And remove R2 47K from output , it is loading the preamp too much and is not present in the original schematic . I think it was stated that the load for this circuit should be no less than 47K and somebody not very smart understood that 🙂
Last edited:
Single point grounding is worthless at 800 MHz! To avoid ground loop , input connectors should be only ground connections - use nylon standoffs if PCB is grounded (you can ground it through a 100 Ohm resistor if you ever have to operate it with rear panel off). Output connectors should be grounded - if shielded, connect only at one end. Sand paint off under I/O connectors or use star washers. 22K grid stopper should not have any audible effect; directly at grid of first stage. if you use bypass caps, they should be close to input or output with minimum lead length.
The case may be a problem since panels don't overlap and don't have enough contact points to be an effective shield - at 800 MHz, 1/2 wavelength is 7.3". If I were designing a box for EMC, all panels would be connected at points at a fraction of that spacing (no inside paint, bronze finger stock to avoid using tons of hardware). Maybe you won't have to go that far... I'd offer you a filtered AC connector but looks like there isn't room. With plastic connector, use solid wire, minimum length, to both mounting screws.
The case may be a problem since panels don't overlap and don't have enough contact points to be an effective shield - at 800 MHz, 1/2 wavelength is 7.3". If I were designing a box for EMC, all panels would be connected at points at a fraction of that spacing (no inside paint, bronze finger stock to avoid using tons of hardware). Maybe you won't have to go that far... I'd offer you a filtered AC connector but looks like there isn't room. With plastic connector, use solid wire, minimum length, to both mounting screws.
Last edited: