Input "hum" can arise from small ground potential differences between signal inputs and outputs.
The grounds between amplifiers are established by the AC outlets, and through signal cables. Various factors can cause small ground potential differences between the sound input (power amp) and output (pre-amplifier).
A one-volt difference, while seemingly insignificant, can cause large loop currents to flow through the ground conductor of the signal cable. Depending on the circuit configuration of the audio hardware, the audio signal impedance and the impedance of the signal cable ground, the currents may cause induced voltages in the audio circuitry.
Source: Channel D
This is a source of the familiar 'hum' problem.
In the “old” days, like the Quad 33 and 303 combination, the 303 has no separated ground connection. Grounding was done trough the signal cable between the 33 and 303.
The advantage of such a “star” system: no hum loops. With modern safety regulations such a star system is highly illegal in most countries, and more important, not safe.
In the multi channel (Stereo) world the problems are even more pregnant. Also in most cases the aerial connection of the tuner is also connected to “ground”. In professional installations signal transport is isolated from ground connections (balanced connections); in normal audio systems engineers mix signal ground and chassis ground.
How to avoid hum loops?
Use only one wall socket for the mains connection, if possible. Use the preamp and the associated cabling in the “star” configuration. Separate the safety ground and the signal ground in the power amp, so the chassis is connected to the safety ground of the wall socket, but not connected to the signal ground of the preamp. The same recipe for other equipment connected to the preamp (not always possible!).
General advice; use cables as short as possible. Use shielded cables for the power cables. In some cases, a so called “isolator” is unavoidable; first test one in the aerial connection of the tuner. The tuner signal is FM modulated, so an isolator does less (no) harm to the signal quality. In RCA cables a small signal transformer play this role. Don’t spend a lot of money on special cables and end up with a cheap isolator in between.
In some cases increasing the resistance of the shielding of the signal cable has a positive effect, apply this only if the amplifiers have a proper installed safety earth connection!
How to detect and solve hum loops?
In general; a hum loop generates hum independent of the volume settings. If the hum is dependent on the volume setting; the hum is equipment borne in most cases. But the hum could also be a combination of both situations! The elimination of a hum loop is a painstaking process. Trial and error and applying logical thinking is the best solution.
Start with only the power amp and the loudspeakers, no signal cables connected. Then rebuild the configuration item for item. A cable is also an item! Once the source of the hum is known, a measure can be determined.
Measures
In a Quad 405 and 303 the chassis ground can be separated from the signal ground. In both amplifiers they are bonded together at the DIN input chassis part. In a 405 with factory fitted RCA inputs the two grounds are already separated. The same applies for a 606 and the higher numbers. In Quad tuners it is difficult to perform this trick, but in most cases the hum loop is caused by the aerial connection. A coaxial cable isolator will fix this problem. For testing and analysis it is acceptable to use a non earthed power extension cable for detecting the source of the hum, but always restore the safety earth connection to the chassis afterwards for your own safety! So you should *either* have 3-core mains leads from every piece of equipment *and* signal cables with the shielding disconnected at *one* end, if separation from the signal ground and chassis ground is not possible, or else a single mains connection and everything earthed via the signal leads, as in the case of a 33. Read also the comments before you change anything to your equipment!
A simple DIY HF isolator!
Joost Plugge
Dear Joost,
ReplyDeleteI've read your article on earth loops and hum with interest, however, the advise to seperate signal earth and compontent earth worries me. In the latest versions of the 405-2 this is already the case, but in the earlier versions of the 405 the signal eart MUST be connected to the device' earth, otherwise you will damage your PCB's!! (I remember R31 starting to smoke...) Please mention this in your article.
kind regards,
Bart Kuperus
(Quad405@hotmail.com)
Pls. give my regards to Stefaan and my compliments to you excelent blog!!
Hello Bart,
ReplyDeleteDave Price had the following comment on 1-1-2008.
First of all, great site I’m looking forward to when your dada site is up and running.
You may already be aware of a PCB layout fault on 405 boards up to and including M12368 Iss. 10. The Zobel network (C12 and R39) returns to signal ground and not power supply ground as it should do. However this is not normally a problem as the DIN input socket ties the two grounds together. If you remove the board to test it or fit isolated RCA phono sockets you effectively connect the output of the amp (via the Zobel) to signal ground and the amp will go into oscillation and R2 will begin to smoke.
The best way to rectify this is either to install the RCA phono sockets in parallel with the DIN socket as you have done, or move the Zobel network to return to power supply ground as all boards after Issue 10 do.'
I hope I haven’t stated the obvious here. It took me ages to track down why R2 kept smoking when i was testing some 405 boards.
Quad introduced the solution for this problem in the Amplifier board M 12565.3, serial 59001 and onwards.
In the situation I described in the “Hum” article, the input ground is still connected to the chassis of the 405, but via the signal cable and both the safety earths of the power leads of the 405 and the preamp. So the situation of oscillation and the rest should not occur, but to be honest, this is a tricky layout! The best way to solve this is to rewire the amp as I described in the article about the floating double Psu for the 405. Reconnect the input ground to the output ground but not to the chassis. This is not an easy solution. Lucky for us ground loops in proper installed Quad installations are rare.
Joost Plugge
Dear Joost,
ReplyDeleteThank you for correcting me and your extensive answer on this issue. Dave Price is right, it is indeed R2, however, I remember R39 (10R) to smoke as well. What I didn't know is that it was a PCB layout goof and corrected in later versions.
As to your remark "ground loops in properly installed Quad installations are rare" I must say that in my experience the 33/303 combo is particularly prone to hum. Especially when you try to use it as a home cinema system and using longer than normal cables. Furthermore, the filters in the 33 introduce some hum when used as well. This can be annoying in combination with high sensitivity speakers or transmission lines.
kind regards,
Bart Kuperus
Hello Bart,
ReplyDeleteA long RCA cable is a hum loop in itself. Remove the outher jacket and connect the two shields could help. When conecting to a home video system, is there also a second antenna connection? The hum in the 33 could be a placement problem, if a strong hum field is nearby the unit, the coils in the filter could pickup hum. Maybe replacemnt will help.
Regards Joost Plugge
BY JOVE I THINK YOU'VE GOT IT
ReplyDeleteI HOPE THAT TEACHES ME NOT TO MESS WITH MY 303 WITHOUT ASKING FIRST
(I CHANGED THE 4 PIN DIN TO 2 X RCA's) AND WONDERED WHY THE HELL IT SMOKED)
THANKS TO ALL
LOUIS
Hello!
ReplyDeleteI can't see dates on your comments so I don't know if I'm super late or not... but...
I am experiencing an odd ground loop. I recently acquired a 405-2 (DIN socket) and 33, which sound lovely. Apart from a ground loop hum.
I took the advice from your post - with the 4-pin DIN plug detached and no input from pre to the 405-2 there is no hum. There is an almost inaudible hiss (ear right up to speaker), which I expected, but no other hum or noise.
When the 4 pin din is connected between 405-2 and 33 (and 33 is not switched on - no power to it) the ground hum reappears
I unsoldered and insulated one end of the ground wire (as per your advice - disconnect one end of the signal ground) in the 4 pin DIN but when connected to the powered-down 33 the hum is still there.
I noticed when I took the sleeve off the DIN plug that there are only 3 connections, L, R and ground. The 'P' pin denoted on the back of the 33 had no connection. I'm struggling to find out what this connection is - can't find any info on the web.
Your help would be very much appreciated!
T
So sorry - totally forgot to mention that the only part of your advice that I've not followed is using the 'star' configuration. The reason why I have the loop is because the 405-2 and 33 are both wired directly to the mains with three pin plugs, hence trying to remove signal ground at one end of the signal cable... unless of course I've misunderstood the solution!
ReplyDeleteHopefully this question might help others in the same situation.
Hello,
ReplyDeleteRemoving the signal shielding connection is not a good idea, in that way you introduce hum. First test with the main earth removed on the power amplifier site.
Joost
I am using the original 4-pin din sockets to conenct my 44 with my 405. The original Quad 4-pin to 4-pin lead did not connect the outer metal part of the plug to the lead. I have found one cheap moulded lead being sold as a suitable lead does in fact connect the outer metal part of the plug - in fact the outer metal parts of the plugs at both ends show continuity. Replacing the lead with one without the outer metal parts being connected resulted in a substantial reduction in hum.
ReplyDeleteDear Stefaan,
ReplyDeleteI am using a 4-pin din socket to connect my quad 405 directly to a macbook pro. Both are connected to the same electricity socket and there is a terrible hum noise independetly from volume setting.
What should I do?
Miguel
Most likely the earth connection is broken.
ReplyDeleteJoost