Weird Noise When Turning Off E-Piano's Amp

Started by sarakisof, December 07, 2022, 11:50:15 PM

Previous topic - Next topic

anotherjim

So, have you decided to add a jack socket to the case?
You probably only need a socket with switching contacts, however, the switching needs to be before the output transformer so it routes there with a plug inserted or routes to the preamp with it out. If I've got that right, you can't use just any switched socket as they do not isolate both circuits from the final output connection. The only commonly available isolated ones I know * are intended to be PCB mounted - they are used in some Fender and Marshall amps.
I do have an example of a traditional metal frame 1/4" socket with an impressive array of contacts with insulated operating, it came from the headphone jack of a Farfisa organ. It would be ideal but I don't think anyone makes anything like that now *.
* I stand to be corrected!

I think Rob may have already suggested the best solution - a relay. With the preamp switched off, the contacts route to the transformer. Switched on, the relay picks and it routes to the preamp.
The lowest-tech/easy way would be to fit a switch by the extra socket to do the same thing.


Rob Strand

#61
Here's the idea:



You will need one of these sockets:
https://www.wiltronics.com.au/product/5170/6-5mm-stereo-socket-with-dpdt-switch/


To explain a a few things:

(a) The whole idea is to add a passive Pianet T connection to a Pianet M unit but without the pickups being loaded by both the transformer or the Pianet amplifier input.

(b) The Pianet amplifier ground is disconnected from the socket ground to prevent hum loops when then Pianet is plugged into the mains and the passive connection is being used.

These added requirements means we need to use a 6.5mm socket with isolated  DPDT switches.  These socket have been readily available for a long time.

The beefs I have with the solution are:
- The signals have to pass through contacts.
- Switching ground is a little undesirable but it meets the requirements.
- Those 6.5mm socket aren't as reliable as the conventional sockets.
   (but just the same, they are used in many products.)
Send:     . .- .-. - .... / - --- / --. --- .-. -
According to the water analogy of electricity, transistor leakage is caused by holes.

sarakisof

#62
Thank you Rob. After Jim's idea about using switch, i was about to use a DPDT switch and just a simple jack socket. The same idea like yours, but using a simple TS socket instead of one with isolated DPDT switching (J1).


Two questions:
1. We want the specific J1 in order to switch ground and prevent hum loops, so i guess using a common one is denied.
I have many open stereo switched sockets in my shelves. Why these won't work?
2. In your diagram the switch is a DPDT one, correct? Or it's a relay? If it's a relay, then if i use a DPDT switch, do i definitely need the J1?

Rob Strand

#63
Quote1. We want the specific J1 in order to switch ground and prevent hum loops, so i guess using a common one is denied.
It's not denied but there is a chance of a hum loop if the Pianet is plugged in (on or off) and the external amplifier is also grounded.   A some point these problems always show up.

Another way, perhaps even a better way around this, is to not switch the ground.   What we do instead is use the transformer to do the ground isolation.

The changes would be:
- Don't switch the Pianet side ground with the contacts. The Pianet grounds always connect inside the unit.
- Break the ground connection between the primary and secondary of the transformer; currently the Pianet-T
connects the ground on both side. That's fine for the Pianet T because it doesn't connect to mains earth.
The primary (Pianet) side of the transformer connection to the Pianet ground, and the second (external) side
of the transformer connects to the socket ground.
- Optional but good for preventing RF issues.  Add a 100nF cap in parallel with 100 ohm between the
  socket side ground and the Pianet side ground.

See figure 2 on this page for the resistor+ capacitor idea, (in your case use 100 ohm)

https://sound-au.com/project35.htm

Treat the left side as Pianet (no sockets) and the right side as the added socket.  No need for the earth switch.

Quote2. In your diagram the switch is a DPDT one, correct? Or it's a relay? If it's a relay, then if i use a DPDT switch, do i definitely need the J1?
No, that socket I linked has a DPDT contact built in.   That's the advantage of those sockets.
Send:     . .- .-. - .... / - --- / --. --- .-. -
According to the water analogy of electricity, transistor leakage is caused by holes.

sarakisof

#64
QuoteNo, that socket I linked has a DPDT contact built in.   That's the advantage of those sockets.
Cannot find those DPDT sockets here.
What if i use a DPDT switch and a mono socket next to it like that?




anotherjim

Think about how you will use the instrument.
As it's decided it can't do both modes at the same time and it must have separate outputs, how do you want it to work in practice?
If it were me, I'd want to have it set up so I could leave things plugged in and only need to switch things on or off. I'd have the M feeding a mixer, but the T's were often used like guitars with pedals and combo amps so the separate output is a good thing.
I'd go with the separate changeover switch.

It will be highly likely both feeds will route to the main safety ground, so follow Rob's suggestions with ground treatment.

The 10R//100nF network shown here really is effective. You don't need the ground lift switch shown.





Rob Strand

Send:     . .- .-. - .... / - --- / --. --- .-. -
According to the water analogy of electricity, transistor leakage is caused by holes.

sarakisof

#67
Quote from: Rob Strand on December 29, 2022, 02:57:15 PM
What about this, simplifies switch wiring,



Nice! It was in front of my eyes but i had my mind into complicated things and couldn't see it. Seems the best and simplest solution so far.
Going this way, we don't bother with ground switching.

The switch in this diagram is a common spdt (on off on) one, correct?

I don't have a 100nF old cap in hand (only chinese films - you know about my love in old components), but have some 82nF, 150nF tasty mustards. Can i use one of those instead (82or150nF) and should i change 10R to another value then?

Rob Strand

QuoteThe switch in this diagram is a common spdt (on off on) one, correct?
Yes,  as simple as it can get.

QuoteI don't have a 100nF old cap in hand (only chinese films - you know about my love in old components), but have some 82nF, 150nF tasty mustards. Can i use one of those instead (82or150nF) and should i change 10R to another value then?
It's not critical at all.   I'd probably go for the 82nF.

The resistor isn't that critical either (10 ohm to 100 ohm).   For stuff like this I'd probably use something around 47 ohm to 100 ohm and only try 10 ohm if I could hear some buzz in passive mode.

FYI, the way the switch is set up it just leaves the amp input floating. I figured a non-inverting input with 1k5 input impedance + connection with a coax is going to be fine.   You could add a 1M to ground across the amp input to stop the pop when you switch from passive to active mode.   Another way is to use another switch contact and short the amp input to ground - I was trying to avoid doing that.
Send:     . .- .-. - .... / - --- / --. --- .-. -
According to the water analogy of electricity, transistor leakage is caused by holes.

sarakisof

I just finally sourced the transformer. Quick tested with alligator clips connected straight to pickups then to a jack and it works indeed. Apart from its four pin legs it has the common metal screening mount. How do i connect the two screening pin legs? To gnd alongside with other two pri. & sec. "gnd" pins, or just solder alone just for supporting?


QuoteYou could add a 1M to ground across the amp input to stop the pop when you switch from passive to active mode.
Like that (red colour edit)?


Rob Strand

Quote
I just finally sourced the transformer. Quick tested with alligator clips connected straight to pickups then to a jack and it works indeed. Apart from its four pin legs it has the common metal screening mount. How do i connect the two screening pin legs? To gnd alongside with other two pri. & sec. "gnd" pins, or just solder alone just for supporting?
It's probably best to connect them to the external jack ground, just use a wire.

QuoteYou could add a 1M to ground across the amp input to stop the pop when you switch from passive to active mode.
Like that (red colour edit)?
Yes, that's it.   It's better if it doesn't pop, more professional.
Send:     . .- .-. - .... / - --- / --. --- .-. -
According to the water analogy of electricity, transistor leakage is caused by holes.

PRR

I don't think that is a "screening mount". I think it is just to hold it to the PCB. I have frequently used these by just gluing them upside-down (scrape any wax coating) to the chassis and bring wires off the leads.
  • SUPPORTER

Rob Strand

QuoteI don't think that is a "screening mount". I think it is just to hold it to the PCB. I have frequently used these by just gluing them upside-down (scrape any wax coating) to the chassis and bring wires off the leads.
Technically not a screen but it does provide some screening, best not to have floating metal.
Send:     . .- .-. - .... / - --- / --. --- .-. -
According to the water analogy of electricity, transistor leakage is caused by holes.

sarakisof

#73
Definitely it's not a common screening can. Have seen them as output trafos in old scrap crt TV's I've salvaged in the past and those pins were just soldered alone to provide stability just hold to PCB.
QuoteIt's probably best to connect them to the external jack ground, just use a wire.
i had done so in my G9 Gyraf tube pre back in the old building days (now only repairs and mods 😂) for input & output OEP trafos (with normal screening cans) with success.

So, i won't solder those "shield" pins into ground plane, just solder alone holding to PCB and only connect one pin to jack ground via wire, correct?

EDIT: What about that 2nF bypass cap across secondaries?
Should i install it normally in the new diagram, right? (blue colour edit)



PRR

  • SUPPORTER

Rob Strand

QuoteEDIT: What about that 2nF bypass cap across secondaries?
Should i install it normally in the new diagram, right? (blue colour edit)
You can spend a lot of time testing to choose that cap.  It's not uncommon for technical reasons to add a resistor in series with the cap.  It also depends on the fine details of the *specific* transformer.    The cap can be used to tune the high frequency performance of the transformer.  However it does more than that it can help reduce parasitic resonances.

The value doesn't look unreasonable to me.   You can only take the manufacturer's recommendation, 2n2 would be the closest value.  If you wanted to be a little cautious you could back it off to 1n5.
Send:     . .- .-. - .... / - --- / --. --- .-. -
According to the water analogy of electricity, transistor leakage is caused by holes.

sarakisof

Happy new year everyone. Back to base after some days off. I'm in the final procces with Rob's diagram.
Just a question. Does polarity of the mini audio transformer matters?





anotherjim

I usually open out stripboard/perfboard holes to take the traffy mounting tags. Bend them over, they are eminently solderable. If the traffy can go at the edge of the board, one tag can hang over and be clenched around. If you can have a board only the width of the traffy, then both tags can be clenched on the edges. And yes, I'd connect a wire to ground.

Some small PCB mount audio transformers I've seen have a mechanical drawing showing the reference "dot" for polarity at opposite corners although, on schematics, the dots usually show at the same top end. Those with coloured wires can simply list the wires.
I can't find data for the pcb mount version you have although if you can find a photo of the wired version, those wires probably come out at the same ends of the body.






sarakisof

#78
Yes, i didn't mentioned that in previous post, i already used this paper note which came up with my traffo bag



and compared with some 45-703's with coloured wires i found online, realised that the polarity is opposite corners order (mine doesn't came up with coloured wires, just plain legs and a "P" marking for primary side and probably at the start winding).





In a simple alligator clips test i made, the result was the same whatever polarity i wired them though.
It worked perfectly like that for example (red and white are hot/signal tips, greens are gnd sleeves , so same direction).



Do you think the polarity is critical here? If i have to wire them in opposite direction it will be difficult in a stripboard board I'm planning to use.
I'm also thinking of building it in a point to point terminal strip board.

So far, my idea is to avoid opening new holes or damaging / altering as less as possible original build.
First idea was to build Rob's diagram (board and traffo) in a board in the keyboard section (upper right, exactly like in the Pianet T) and open two holes (switch and jack) underneath wood base (holes like pickups cavities so you can have hand access to release toggle switch and in order toggle and jack wouldn't prevent instrument lying flat on table).
So the mod would be made into wooden keyboard section only, no amp cabinet upper case damage , it would be playable (rca amp disconnected) without amp only via keyboard itself, exactly just like in the T, very practical for service/troubleshooting purposes etc.. No one could notice any holes / damage until he would look underneath.
It was the perfect idea, until I realised the M has stand legs, it's a spinet instrument and that the only free space i was planning to drill those two holes, was the place where the legs and the upper amp cabinet are mounted to the keyboard.   :icon_cry:

So, the idea goes to amp cabinet, especially in the power socket / outputs metal panel area.

"Hmm, i can see two already drilled round DIN holes, headphone and extension to hi fi speakers outputs I'm probably never going to use..."
After days of thinking, came up with this. In order altering originality as less as possible, i removed the rivets of the two outs without damaging them. They were ready to come off prone to years of usage anyway, so screws instead of rivets are always a good thing in those cases if someone wants to remove my mod and put the original DINs back in the future.
So DINs will be still there mounted inside with cable adh ties. The user can use them if open the panel only. Will never use them anyway.

In DINs place i will install the switch and the jack. Found some creative ways to mount them as their diameter is smaller than DINs hole. Would be better if i could find a more thick metal/aluminum even plastic plate though. Those are those thin "on off" name plates from older switches.



And the final result from the outer side will be smthng like that (with screws and maybe a blueing / darkening of the plates / or other plates will be even more beautiful). No extra screws, no any new hole opened. The unit is like before.



Now the big thing is about the small circuit/traffo combo. Where exactly and how? Two days thinking of many ideas but cannot decide yet. 😂
Best bet would be a double sided PCB board (Eagle and PCBway) in the back of the panel , so the switch and jack would be soldered straight to pcb , and mounted everything there. Like we do in modern stomps. Jack and switch hold the PCB itself. Of course traffo and components all soldered in that same pcb. Simple and end of story. But this would be so modernized for my tastes and Pianet's nature. Would like a fat heavy duty toggle and cannot find one with pcb legs. Pcb legs are used in those small thin toggle ones. Slide switch would be good (they were widely used in the past too, see Philicorda organs underneath so they can lie flat), but then i have to find a metal/aluminum panel with rectangular slot/notch and those switch nameplates have round hole indeed. Damn mechanical things again as always..😡

I'm thinking smthng like a p2p layout for traffo, 10ohm,100nF, 1M, 2nF or completely handwired in the air, or on a strip board. Maybe i can mount the board on the already drilled screws..
Any ideas welcome..




anotherjim

I don't think you need to worry about polarity in this case. If you were to use the T & M outputs together, it would be best if the phases matched, but you won't be. You can do whatever is easiest to lay out. It is, however, best practice to maintain original phase through a circuit if you can. It won't matter if the start ends are grounds.

The way Farfisa added stuff was to install an aluminium angle for the jacks and switches mounted so you had access through a cutout in the rear panel. Slide switches would be used for their low profile.
A PCB would be provided, simply screwed onto the base with plastic plain standoffs.  Often, wiring pins were fitted to the board as the way they were made was to complete the mechanical assembly and then fit prepared cabling in situ.