Syntrillium Backup Forums :: View topic

AndyH · Posted - Fri Jul 12, 2002 4:03 pm

There are certain aspects of this hum from the turntable about which I would like to know more, if someone is feeling generous. Besides what I already listed in the 7/10 posting, if I disconnect the turntable's grounding wire from the receiver's case, there is a BIG increase in hum and noise. This is perhaps to be expected. If I instead remove the headshell, leaving the two audio channels open (not a complete circuit), there is also a big increase. Is this latter to be likewise expected, or might it be a clue to anything relevant?

I mentioned that there is no difference in the hum whether the turntable is turned off or turned on. There is also no difference if the motor is running (direct drive platter) or not running. Furthermore, there is no difference if the turntable is unplugged from the mains supply.

I have the idea that ground loops are caused by two or more connections to ground that are not at exactly the same potential. There is a current flow across the resistence between those points, and probably towards the ultimate lowest ground. This generates the hum. The cartridge is the source of voltage under normal operation, and thus of the very small current flow to the phono preamp input. The phono preamp does not "energize" the cartridge, and its associated circuitry, does it? If the turntable is disconnected from power, where does the current flow from and to?

Another aspect: Putting the power supply in one case, and a preamp (not necessarily a phono preamp, consider no turntable involved here) in another case is, as discussed in a different thread, the surest route to preventing induced hum in the preamp. While both circuits are enclosed in metal cases, there are jacks, holes for controls, etc. so the circuits are not in complete Faraday cages, but they can be somewhat reasonably shielded from outside influences, no?

Suppose everything is designed properly and constructed well and working swimmingly. Now change the preamp by removing several ground connections from the star configuration or buss bar, and attaching them instead to the aluminum chassis at various points. We have thus created one or more ground loops and considerable hum will probably appear at the preamp output. However, nothing but well filtered DC goes into the preamp box. I could understand broadband noise, such as any resistor generates, but whence comes the 60Hz (or 50Hz across the wide sea)?

SteveG · Location: United Kingdom

Ah, but how much does it hum if you remove the supply connection completely? Let's look at the potential coupling mechanisms.

One of the functions of the ground lead is to prevent the turntable chassis from re-radiating hum that comes in on your supply cables, which is then capacitively coupled to it. When you consider the relative levels of signals that we're talking about here, and also remember that the RIAA preamp has a pretty large boost around the 50/60Hz mark, this is not so surprising, perhaps. Also, an ungrounded turntable can have hum coupled in by the same mechanism from room wiring, etc.

The other thing that you tend to notice with ungrounded T/Ts is that the hum level increases when you start the motor, if it's an AC synchronous type, because hum then gets induced into the base plate electromagnetically, probably at an even higher level...

How much hum does it take? If you have a 110v supply, then just coupling 0.00001% of this into the cartridge or leads is going to be loud!. And although it's no real comfort, just bear in mind that in the UK, the problem is over twice as bad - we have a nominal 240v supply. If you can get hum and noise to be at about the same level, you've done moderately well. If you can get hum levels significantly lower than this, you've done a pretty good job.

But reducing these last vestiges of it is undoubtably the hardest part!

Steve

SteveG · Location: United Kingdom

I should perhaps add that these mechanisms work universally, so all of the same things apply to preamp boxes, etc. That's why it is common practice to put potential electromagnetic radiators like transformers outside the box, or at least have rotatable fixings (very common at one time) to align the transformer for minimum hum induction.

Steve

Edited by - SteveG on 07/13/2002 03:37:51 AM

Havoc · Posted - Sat Jul 13, 2002 4:13 am

As for your specific questions:

An open channel can have more noise because the impedance of the circuit is higher (unless the cartridge is the source of the problem). If your input is loaded with a cartridge, the cartridge determines the impedance of the loop (one end the loading resistance in the preamp, about 47k, other end the cartridge, about 1k in between the cables). So the current has to folow in a 1k loop or is induced on the 47k resistance. For a give amplitude over a resistance, you need far less current in the 47k loop. Now mostly, hum is induced through a high impedance, like capacitively from a nearby mains. So, you get a voltage divider, one end is always high resistance (coupling from mains) and the other end is either high resistance (open leads) or low resistance (cartridge). The high resistance case gives the most voltage at the input of your preamp and so the loudest noise.

The current flows because it is inductively coupled into a closed loop, or it is capacitively coupled into a conductor. So it comes from the coupled mains and looks for a way back to the mains earth. This can be by a real mains earth connection, or even trough the capacitive coupling of the windings of your transformer.

A faraday cage is great, but do not overestimate it. For capacitive shielding, it is efficient, for magnetic shielding at 50-60Hz, you need lots of material. Think a couple of mm steel. This is why getting your transformer away is a good idea for that problem, nothing beats separation. With connectors, it get more complicated, currents follow the way of least impedance, not always the one you want (or think they take). So if the 50Hz has a lower impedance by following the signal into the box, it will do so and can then couple inside the box capacitively into your circuits. In reality, the skindepth plays a big role in this, but this should give you an idea what happens. This is why the ground lead makes such a big difference, it provides a low impedance path that is controlled outside the circuits.

To get a idea of a real preamp, imagine capacitors (they are there really) between every wire going in and out of your box to the mains/earth connections in your room and also to the box itself. The small capacitors couple mains directly into everything. So any imbalance between them causes currents to flow. These are small, but some audio circuits are very sensitive. Now, your example of the star-distributed ground can be valid, but it does not have to be! All depends on the size of the loops relative to the frequency, the impedance of the loops and how good your ground plane is (this is what you create). It can well be that by doing this, you make a bypass for some current, and that this solves a ground problem.

With hum and ground related problems, theory is one thing, but lots of experiments are almost always needed.

SteveG · Location: United Kingdom

AndyH · Posted - Wed Jul 17, 2002 3:23 pm

and still more ?? I have read a little on grounds and hum, probably the proverbial enough to make me dangerous to myself.

An article explained a common cause of ground loops in regular equipment (not screwy home brew with crossed wires). This involved different potentials in the safety ground (the third prong on the plug) of various pieces of equiptment, especially when plugged into different mains outlets.

Looking at my audio and video equiptment, all racked together, I see that not a one has a three prong plug, which must mean they are all "double insulated." Its kind of difficult to tell some things, from my cursory preliminary examination, since the cases seem to be insulated by the paint on their outside surfaces. For the receiver, where every thing else meets:

Should the "ground" side of all those RCA jacks be (electrically) connected to the case (vs just to the circuit inside the case)?

Should the receiver's turntable grounding screw be connected to the case?

The article briefly mentioned doubled insulated equiptment, mainly to say that its message did not apply to such stuff. One thing about that type of equipment it did say is that an earth connection is not needed, and "in some cases must not be used." I presume the "must not" is for safety reasons. Is it possible, barring some insulation failure that might make things dangerous, that a sound quatity improvement might be made by putting a grounding lug onto the receiver's case and running a wire to ground?

Building a preamp may create a potential hum problem. I would not expect to be able to duplicate double insulation standards (and I have little idea how such things are constructed anyway). The reasonable thing, safety wise, would be a three prong plug with the case at earth ground.

If the rest of the equiptment is double insulated, is there likely to be more of a hum problem to overcome than there would be if this new preamp were also double insulated?

Is so, should some particular construction technique be followed, such as not connecting the signal ground to the case? (since the preamp cannot be doubled insulated)

Would this require that the output side of the power transformer, and everything thereafter, not go to the case either? Would this not create a safety hazard, in the rare event of transformer short out?

When I originally ask the question about hum with the turntable plugged into the receiver vs no hum when the turntable was removed, I wanted to know if that necessarily meant a grounding problem. The answer was yes. Now, reading this thread's info about induced hum, I ask again.

Since there could be hum induced capacitievly or magnetically into the cartridge circuit, and since the circuit is unbalanced (thus no common mode rejection) might there not be some significant level of hum injected into the preamp from the turntable, even with perfect grounding, or am I just not understanding enough?

Edited by - AndyH on 07/17/2002 3:26:48 PM

SteveG · Location: United Kingdom

If everything really is floating, then this is probably quite good from an interconnection POV, with the possible exception of the T/T, as explained before. Yes, you should have a common ground area on the back plate of your reciever. Normally, the T/T grounding screw is connected to this as well, and this can only really cause problems if there is a separate ground for the motor drive. You may have to check this out. If the T/T chassis isn't grounded, and there's a separate ground lead to connect it to the reciever, then the best approach is to use this, and then have one common ground for the reciever, which it sounds like is missing on your system.

So briefly, you star-earth each individual item back to the reciever chassis, and you ground that back to one ground point (get a three-pin plug and use the ground from here). No other grounds allowed! If you have a T/T 'complication', then you can ground that separately, but then you have to isolate the arm, as we went through before.

As for the last question - well, it gets complicated. Essentially in practical terms, it will always be possible to inject hum at some point if you really make the effort, and ultimately you will be limited by the practical possibilities. If you've understood the problem, the next thing to do is to experiment and see what happens. If you start with the scheme above, that will be good. But when you come to connect a PC (which is usually grounded) to the system, you may have to rethink it slightly. I have a rig here in which the PC and preamp share a common ground quite happily with no hum problems at all. Everything else (including the power amps) is returned to the preamp chassis, including the Thorens T/T, but this only has a 2-wire supply cord. That doesn't make it double-insulated, it just means that the ground has to be arranged separately, which it is via the large screw termnal on the preamp.

Other than that, the other bits of kit are grounded via their signal leads only (they are double-isolated). Even with loops in and out for the MD, DAT, cassette and open-reel machines, this doesn't cause any hum problems. Basically, once you've got the T/T wiring correct, everything else just falls into place!

Steve

Havoc · Posted - Thu Jul 18, 2002 3:34 am

Paint does not count for double insulation! The reason connecting to earth may be dangerous is that in some places neutral is used for this.

If you can, use an optical connection to the pc. One less possible loop.

SteveG · Location: United Kingdom

Havoc · Posted - Thu Jul 18, 2002 9:39 am

Not a receiver but an external AD/DA with optical IO is not that difficult to find. Or a coax with transformer coupling.

Twang · Posted - Thu Jul 18, 2002 12:32 pm

SteveG · Location: United Kingdom

AndyH · Posted - Thu Jul 18, 2002 5:47 pm

This will require a few re-readings, and then who knows what else, to be clear.

I was not under the impression that the chassis paint is part of the double insulation. The paint only prevents me from using an ohm meter to determine if the TT grounding screw, and the "ground" side of the RCA jacks, are directly connected to the chassis. I'm sure I can eventually get around that barrier, by opening up the case if nothing else. That would require first disconnecting everything, so I am not intensely eager to do it. Directly connected to the chassis seemed likely, but I am not familiar with the details of "double insulated." If I read the replies correctly, those things are currently connected to the case.

The article I mentioned explained an "always safe" (but not necessarily in compliance with local electrical codes) way to break ground loops that can exist with the third prong safety earth. To apply this fix, both sides of the audio circuit must be clear of the case. This lead to the obvious (with my limited knowledge) speculation that the same condition might exist in double insulated equiptment. Your answer seems to be "No."

This article also mentioned, in passing, that unsafe conditions could sometimes be created by connection the case of double insulated equipment to the safety ground circuit (such as running a wire between the case and the third prong at the wall socket, I guess). It did not explain anything about how to recognize that potential. SteveG does not consider that safety hazzard a real potential, since he suggests grounding the receiver's case to the safety earth.

The article's kind of ground loop problem can only exist if there is a potential difference between the safety grounds of two or more pieces of equiptment. Ideally that condition never exists, but reality is not always ideal. It is exceedingly unlikely to be part of the situation in my current setup, but if I build a preamp, such as the SSM2018T circuit discussed in the other thread, I should probably use a safety ground on the power supply. This will introduce the possibility of more hum mischief, that would not exist if I made it with only a two prong plug.

AndyH · Posted - Thu Jul 18, 2002 6:22 pm

I hope this isn't disgusting annoying. You no doubt feel that you have answered the question about the TT twice already.

The TT ground wire goes to the receiver case. Connecting that wire vs disconnecting it makes a BIG difference. Therefore the receiver end seems OK. However, entirely removing the TT from the receiver significantly reduces the hum in the receiver, so that seems to indicate some problem in the TT.

The hum put into the receiver's phono stage by the TT peaks at about -74dB -- it is very much lower than the hum caused by disconnecting the TT grounding wire from the receiver. (again) however, that -74dB peak is a narrow hump in the Spectrum Analyzer that is about 35dB higher than the noise level +/- 15Hz on either side of 60Hz. That is, the hum is fairly low in absolute terms, but it is quite high compared to the preamp background noise.

Since the TT seems to be adequately grounded at the receiver side, the cause of the hum must be at the TT end. It could be one of three things:
(1) the TT designer's did not know what they were doing.
(2) something has gone wrong with the wiring in the TT
(3) the hum is being magnetically or capacitievly induced, independent of any poor wiring, from the TT motor, the motor's quartz PLL circuit, nearby equipment, or the great world at large (since the 60Hz only decreases by 2dB when I unplug the TT from the wall, I guess the induction idea isn't too real, but I am floating in ignorance, so I feel still unsure).

I am trying to understand if the conditions I see DO indicate a wiring problem or, considering the measured levels, possibly just the way this TT is at its best. The form of the symptoms indicate a wiring problem; does the level of hum seem too much for anything but bad wiring?

AndyH · Posted - Thu Jul 18, 2002 6:37 pm

I don't know if this is relevant to the problem, something to be expected, or really odd. I probably should wait till I get home and get more exact figures, but in case it gives some insight that might otherwise be lacking.

Taking a noise floor recording of the phone section, we see something in Frequency Analysis that is not at all level. The noise gets generally higher in level as the frequency gets lower. I believe this is quite normal, it is due to the phono equalization.

The shape of the curve is very similar with the TT plugged into the phone section or without the TT, except for the narrow peak at 60Hz (and some smaller peaks at higher harmonics). However, the level of that curve at every point is different.

When I unplug the TT, the 60Hz (and harmonics) peak goes away, but the noise level at every other point of the curve is higher. The difference is significantly more at lower frequencies than at high frequencies. Below 60Hz, noise level is about 24dB higher with no TT.

Havoc · Posted - Thu Jul 18, 2002 11:25 pm

Andy, your description of the problem is identical to what I have with the lenco setup at home. The noisefloor is not level due to the riaa, and then you get a peak at 50 (in my case) and a few harmonics.

Have you checked the cartridge? Normally there is a ground connection from the cartridge housing to one of the electrical ground connections. For an unbalaced system, it should be connected. But if the cartridge body makes contact with the ground through its screws amd the arm connection AND the ground strap does this also, you get a loop. cartridge body-arm-arm ground-central ground point-arm wire-cartridge ground.

It is a pain, but what is the result is you disconnect the headshell (but leave the cartridge connected)? This will not make a difference if your headshell is nonmetallic.

AndyH · Posted - Fri Jul 19, 2002 12:17 am

The cardridge cannot be connected without the headshell. The headshell plugs into the arm. There are four little pins on the headshell (recessed well over an inch inside a cylinder that is well under 1/2 inch in diameter) and four pin recepticles on the arm. I guess I could do something by figuring how to connect small wires to the cardridge pins and sticking the wires into the arm holes. I doubt that would make for secure contacts.

The headshell slides onto the arm. The arm slides into that 11/2 inch cylinder, with the pins at the bottom. a set screw holds the headshell to the arm. That certainly could be a problem point. I intend to find a product called "Silver Conductive Grease" and try a little on that mating surface.

SteveG · Location: United Kingdom

Beware! It isn't necessarily the T/T that's 'causing' the hum, although that's not entirely out of the question, obviously. It's usually caused by the interaction of the balance potentials between the T/T cables and the potentials between the ground post and the outers of the input sockets on the receiver, and the relationship between all of this and the surroundings. That's what makes tracking down the vestiges of it such a problem. In any given case there isn't a absolutely coorect textbook answer that's guaranteed to work, because all installations are made up of real-world components in the real world!

I can promise you one thing, though; All of these systems are obeying the Laws of Physics. Your problem is to establish exactly which laws they are following in your own particular case, and none of us can actually tell you that without being there. So, quite correctly, you are trying to understand what the real nature of the problem is, and that's actually the clue here. It's only when you've taken account of all of the possible mechanisms and balanced one thing out against another that you'll come up with the best possible fix. But inevitably, this will be a compromise - in a system where you've balanced out a problem like this, things inevitably change. Even moving leads around can sometimes make a difference, if the system's not using balanced interconnects, which most T/T systems aren't.

Havoc has reiterated some good advice about the cartridge connections and grounding at the headshell, which I mentioned in a previous thread, and this seemingly innocuous connection can sometimes cause untold aggravation, because in general terms, it really is in the wrong place! If you work on the basis that you want a single ground connection following through from the ground plane at the back of the reciever to the cartridge, and no dual paths anywhere, then this cartridge ground connection is putting a dual path back just where you really don't want it. It relies on the T/T chassis being at exactly the same potential as the far end of the signal lead ground, but via a different path, and also having the same level of induced supply hum. No chance! It's definitely worth checking this one out.

Steve

Havoc · Posted - Sat Jul 20, 2002 5:37 am

What I mean is: normally there are 4 little wires with contacts that run from the headshell contacts (that go to the arm) to the cartridge. So, you disconnect the cartridge from the headshell without disconnecting those wires. Okay, it means setting up the cartridge again afterward, but it may give a clue as to where your ground loop is situated.

In my case, it gave no improvement. Only when I made a connection not going through the arm did it improved, and quite spectacular. So at the moment I'm rewiring the arm. Took it apart and as there is no way to remove the connector between arm and shell, I'm drilling it out so I can run 1 wire from cartridge to amp.

This reminds me there is something else you can try: take some light wire like wire-wrap stuff, make a long enough twisted pair with it and connect that straight from the cartridge to the preamp, no other connections. See what that gives. Then connect the ground wire between TT and preamp.
Other thing, run an extra ground wire from the central TT ground to the shell. It may be a broken ground wire to the arm.

BE CAREFULL, but you just have to try until you find something that gives a remarkable improvement, and then work your way back.

As Steve said, the problem is the same, but there is no single solution. In one of Texas Instrument's handbooks, there was a chapter about grounding: the last entry for solving them was PRAY. So even they do not know all the answers.

SteveG · Location: United Kingdom

AndyH · Posted - Mon Jul 22, 2002 3:08 pm

I believe I finally understood the idea. However, what happens with non-metallic cartridge bodies and/or non-metallic arms? Why does Havoc recommend a twisted pair from cartridge to preamp?

I tried an easy approach, I can't figure out how to get into the turntable internals without damaging the arm. I simply ran a test lead from a cartridge mounting bolt to the preamp's TT grounding screw. Actually, I first disconnected everything and moved the TT, receiver and computer to another room, so I could be free of other equipment influences, and would have enough room to get at what I wanted. Connecting the test lead wasn't perfect, but it reduced the 60Hz peak, as viewed in Frequency Analysis, by 15dB. It made little or no difference whether or not the TT grounding wire was also attached to the preamp.

After many measurements, I reinstalled the equipment carefully on its normal rack and re-tried the test lead measurement. It made no difference what-so-ever (noise and hum were actually a bit higher than they had ever been before). I hauled everything back to the dining room and tried again. I could not verify the earlier results.

I have the recordings from the experiments, but I am not sure what to believe. Maybe I should burn some incense.

SteveG · Location: United Kingdom

It's very difficult when you want this to be a precise science and it obstinately refuses to be one!

The reason for the twisting (this is quite normal) is to make sure that whatever influences are present have an equal effect on both cables. Obviously, this is only going to be significant in a balanced system.

With non-metallic arms, you rely on either the screening in an unbalancd cable, or pretty good cancellation in a balanced one. But there's no real problem with putting a screen around a balanced pair either. It usually works fine as long as at the cartridge end, the screening is not connected to either pickup coil, which will immediately make things potentially worse by forming a groundloop... you just leave it unconnected, in the balanced case but connected to one side of the coils in the unbalanced state.

Steve

AndyH · Posted - Mon Jul 22, 2002 6:18 pm

Havoc's suggestion for a twisted pair was for a ground wire -- something running from cartridge to preamp, not part of the signal circuit. I don't understand how that would be any different than a single wire, aside from the necessity to assure that both wires were equally well connected.

I have never previously heard of grounding the cartridge itself; I don't remember any mention of such an idea in any of the setup articles I have read. If the cartridge body is not metal, the only way there could be anything to ground would be if there were an screen inside the cartridge. This would require either a fifth wire or for the screen to be equally connected to the ground side of both channels inside the cartridge.

Since my cartridge body is, at least partly, metal, it must be intended to be grounded through the tonearm, if grounding is a consideration. As I described my experiences, a grounding at first seemed to be helpful, but later showed no effect. I guess all of this is one of the things contributing to the high cost of some tonearms.

Reading Havoc's 7/20 posting, I see I might have incorrectly assumed the "light wire like wire-wrap stuff, make a long enough twisted pair with it" was about a ground wire. Perhaps he was suggesting that I bypass the tonearm signal carrying wires.

Edited by - AndyH on 07/22/2002 6:32:36 PM

SteveG · Location: United Kingdom

Well if he's not sending a balanced signal down these twisted wires, there certainly isn't a lot of point... But if it's single ended, then the ground wire will be either part of the circuit, or if it's an extra ground, possibly a part of the problem...

A lot of cartridges have a link from one pin to the metalwork on them (unless of course they have no metalwork), and I believe that there are plenty aroound with internal screening as well. It does make a lot of sense to have screening here, because this is the most likely place for induced hum to be picked up, and it's the one place in which the user can't actually do anything about it. But nevertheless, grounding the headshell by a link from a signal pin is usually just plain wrong.

Steve

Havoc · Posted - Tue Jul 23, 2002 5:26 am

The use of the twisted pair (TP) was indeed to bypass the arm wiring. No, it is not very usefull when not balanced, but it is better than connecting 2 loose wires. I would use that also before trying to connect some RG59 to a cartridge. Wire-wrap wire is very convenient for this kind of trials because it is light, flexible but not to much , easy to find and cheap.

The idea is to bypass the arm wiring, and the tp is as Steve said for equal influence on both wires (even in un-balaced systems). Your signal and its signal-ground from the cartridge will use the tp. The other cable that runs from your receiver ground connection to the TT central ground would be larger. The idea it to create 2 seperate paths that come together at the receiver forming a star ground.

The more I think about this, the more I get convinced this is more a capacitive coupling problem than an inductive coupling problem.

AndyH · Posted - Tue Jul 23, 2002 4:10 pm

So you actually meant to say use two twisted pairs, one for each channel, or perhaps you meant 'bypass one channel at a time with a single twisted pair', testing for an improvement piecemeal?

Words are so slippery. I think we have still been talking about different things. You started off telling me "without disconnecting those wires." Now I think I understand that your twisted pair instructions (two paragraphs after the 'don't disconnect' admonition) implied that those same four wires must now be disconnected, or at least that one pair of them must be disconnected at a time. Am I possibly getting on the right track now?

How did you make a secure enough connection to the cartridge pins to assure that a bad connection there would not be part of the problem?

Havoc · Posted - Wed Jul 24, 2002 5:29 am

Yes, in describing those things is not easy to get "crossed wires". So, a bit of clarification:
- well, 2 pairs or 1 pair does not matter. The problem is in both channels (it is, isn't it?) so you can study one channel at a time.
- I described 2 experiments. In the first, you leave the cartridge connected to the 4 arm wires, but loosen the screws to the headshell. If your headshell and cartridge are metal, then by doing this you break a possible ground loop. It goes like this: 1) your arm/headshell/cartridge-outside is connected to the central ground point. 2) your cartridge metal can be grounded by the ground-strap and the arm wires. So in disconnecting the screws from your cartridge to the headshell, you can break the loop. If this helps, then you know where your loop is and you can start thinking of solutions.
The second experiment is leaving the cartridge screwed to the headshell, but disconnecting the 4 wires through the arm and wiring the cartridge with a loose wire (the wire-wrap twisted pair episode) to the preamp. You can use 1 twisted pair testing 1 channel (and swap channels), or 2 pairs to test both channels at the same time, up to you. I go for the least work, 1 channel, it is only to check a possible cause, not a definitive solution. It is always possible, that there is a broken groundwire to the arm, or that the insulation of the arm wiring is broken, making an unvisible extra ground point and thus creating a loop. This experiment is a sure way to create a star ground from preamp to cartridge.

I used some connectors for pcb pins. Can be found at an electronics hardware shop.