8mm Forum: Aux. Audio Outputs

This is topic Aux. Audio Outputs in forum 8mm Forum at 8mm Forum.

To visit this topic, use this URL:
https://8mmforum.film-tech.com/cgi-bin/ubb/ultimatebb.cgi?ubb=get_topic;f=1;t=004966

Posted by Steve Klare (Member # 12) on November 11, 2009, 08:45 AM:

I'm working on a circuit to take the aux. outputs of my projectors, subtract out the hum and then send the signal to an external amp.

This is a lot simpler if the signal I am taking in is at a fixed level.

Question: is the level of the aux. output on Elmo projectors typically independent of the volume knob? How about Eumigs?

Posted by Martin Jones (Member # 1163) on November 11, 2009, 08:52 AM:

Steve,
Certainly is on the ST600; and industry standards for audio usually expect a level not affected by controls, either volume or tone.All the Eumigs I have diagrams for conform to this.
Caution however with Elmo; their "line out" levels are usually lower than industry standard (and lower than other manufacturers).
Regards,
Martin

Posted by Damien Taylor (Member # 1337) on November 11, 2009, 10:49 AM:

I have an ST-600. The aux out level IS directly tied to the volume control. Some say the best place to 'tap' in to the amp for aux out is directly before it hits the volume control.

Posted by Martin Jones (Member # 1163) on November 11, 2009, 11:15 AM:

An apology is due here! Damien is quite right, the "AUX OUT" socket is at the output of the machine, after even the power amplifier. My confusion arose because in the UK we refer to "Line In" and "Line Out" sockets, which are placed before the volume control. I have such on my ST600 because I actually converted the "AUX IN " socket(which is before the volume control) to act as both "Line In " and "Line Out" (but not both at the same time!) when I converted the machine to Stereo.

I think that the more complex Elmos have true "Line" level sockets, but Kevin Faulkner can probably confirm that..
Again apologies... my bad memory again!

Martin

Posted by Steve Klare (Member # 12) on November 12, 2009, 01:17 PM:

Thanks for the responses, Gents.

As I said, what I'm up to is easier at a constant volume (and therefore hum) level, but not impossible without it being constant.

As Mel Blanc said many times: "Back to the old drawing board!"

Posted by Greg Marshall (Member # 1268) on November 12, 2009, 11:49 PM:

Steve.... have you been sucessful at reducing or removing the hum?

Posted by Steve Klare (Member # 12) on November 13, 2009, 08:01 AM:

Not so far, but I have an idea I want to try. Right now I'm just trying to figure out the ground rules for designing the circuit.

I've already brought a projector to work and put an oscilloscope on the speaker (-during lunch, office door closed!) so I am getting somewhere.

Posted by Martin Jones (Member # 1163) on November 13, 2009, 08:19 AM:

Steve,
Try this....
http://www.radio-electronics.com/info/circuits/opamp_notch_filter/opamp_notch_filter.p hp

One IC will provide either TWO stages of filtering at 60 or 120 hz OR one at 60hz AND one at 120hz for EACH of two stereo channels.

"Line Out" signals should be taken off at the "top" end(s) of the Volume Control(s)if such sockets are not already fitted.

Martin

Posted by Paul Adsett (Member # 25) on November 13, 2009, 08:56 AM:

Steve,
A lot of people, like myself, connect the AUX out to a 10 band stereo equalizer unit first, and then ouput the equalizer to the stereo amp. You can then set the filter levels at 60hz and/or 120hz (or at any of the 10 bands) to filter out the hum level. You can also use the equalizer to boost high frequency levels on 'muddy' sound tracks and also to filter out hiss and low frequency rumble.
I use a Yamaha equalizer, which you can get from Radio Shack or Best Buy for about $120.00.

Posted by Martin Jones (Member # 1163) on November 13, 2009, 09:16 AM:

Paul,
I suggested that particular approach because most commercial Graphic Equalisers can only offer a maximum of about 12dB attenuation at the centre frequency of each band, which tends to be 62 and 125 hz in the States. The circuit I posted offers about 45dB at a set frequency fairly easily, and up to 60dB with careful selection of components and tuning. The notch is VERY narrow, so nearby frequencies are not affected, and more than one filter can be cascaded for even greater rejection.
Component cost for TWO channels, each with two filters at spot frequencies... less than $20 (at UK prices).

Of course, that does not in any way preclude the use of a multi-band equaliser as well to fine tune the rest of the audio range; neither the equaliser or the notch filter will affect each other's performance in any way.
Martin

Posted by Steve Klare (Member # 12) on November 13, 2009, 12:01 PM:

I’ve been chasing this problem around for years, but being that I’ve never really understood what the problem is, I’ve never been able to get a handle on it.

About 5 years ago I built an interface circuit to take the output from the projector and send it into a spare input of my stereo amp. The hum was awful, so much that It sounded better out of the built in speaker.

I added a capacitor to my interface, figuring if I could roll off that 60 Hz. it wouldn’t be so obnoxious. It didn’t make any difference I could hear, so I stuck in more capacitance to raise that corner frequency, and It worked…

But…

Now the sound, especially music, sounded awful. The bass tones were really thin and the music became tinny.

Back to the internal speaker.

A few weeks ago I wound up keeping an oscilloscope in my office so nobody would “borrow” it. I saw an opportunity and brought my Elmo ST-800 to work and set it up during lunch and got this waveform from the audio.

This is with the volume knob halfway up, the lamp lit and the film transport forward and 24 FPS. The fuzz on the waveform is due to the motor and goes away when you stop the transport. There is no film in the machine, therefore no "sound". This is what hum really looks like.

Obviously there is more happening here than line frequency. Since I graduated from engineering school at roughly half my present age, I feel lucky to even be able to spell “Fourier” (...is this right?) at this point. However I am good with Microsoft Excel, so I started adding up sine waves of multiples of a frequency until I got this picture:

I got from this that for this 60 Hz. hum there are three important frequencies at play: 60,120,180 and the ratio of their respective voltages are 10:4:6

Being that the 60 hz is about twice as big as the other two and therefore about 4 times the power, it seems obvious that a lot of progress can be made by just killing the 60.

A little simulation seemed the right thing to do, so I found a software based tone generator and made it produce the three harmonics in the same proportions, lo and behold: artificial hum! It sounded exactly like the hum out of my Elmo.

So I cranked down the amplitude of the 60 Hz. tone…. and nothing happened. It seems that a pure 60 Hz sinusoidal tone isn’t very obtrusive, especially if you have hearing like mine. However 120, and even moreso 180 are obnoxious. It’s not so much a matter of the power that the machine puts out, it’s the power that actually gets all the way out the speaker, through the air, into your ear, into your nervous system and then to your brain that is important. (A friend of mine solved the whole problem by deciding he likes this sound!) This is why my early experiment with rolling off 60 Hz. failed so miserably. It is also why to kill the hum I killed the bass: the real frequencies of interest are much higher than line frequency and are well into normal audio tones.

-now we are getting somewhere.

So to get rid of the hum we need to kill off only the important harmonics and allow desirable sound of the same frequency to pass through unmolested.

What I have in mind is a circuit that isolates one hum frequency and develops a sine wave of opposite phase and fixed at exactly the hum voltage to be used to cancel the hum only. If a musical note of the same frequency comes through, the canceling voltage will stay the same and the note will go right past unchanged. One of these would be needed for each frequency of interest.

-This is hum-bucking, but done electronically, not magnetically.

Potential Problems:
1) The hum voltage changes with volume level
2) The hum voltage may drift as the machine warms up
3) The spectrum of the hum is different between machines (I think I hear a higher harmonic from my Eumig.)
4) Whatever I don’t know yet…(There’s always something!)

It’s not right to say I’ve solved this problem, but maybe I’m beginning to understand it!

Posted by Rick Skowronek (Member # 385) on November 13, 2009, 12:50 PM:

Hi Guys,

Been awhile. Hope everyone is doing well.

Just a side note from my years dealing with high performance audio and tape systems, causes of hum can be a dog to find at times. Typically, this is caused by two main reasons. The first is open or ratty shielding or grounds in the audio path. The closer to the input (like the tape heads) the more the hum and noise. Second, is the power supply. That typically is coaused by failing electrolytic filter caps or even marginal regulator circuits.

To check out further where it's coming from Steve, since you have a good Oscope, is to put the probe on the DC supply in various spots, i.e., input circuits, motor supply voltage, etc. The DC noise and hum should be almost non-existant especially in the audio circuits. If you're getting hum there it will show up everywhere.

To check shielding or tape head problems you may wish to short out the tape head input(s) to see if that lessens the hum. From looking at your supplied trace, I have to strongly guess the power supply and filters are at least partially responsible for your noise.

Hope this helps a little.

Rick

Posted by Martin Jones (Member # 1163) on November 13, 2009, 01:00 PM:

Steve,
Your analysis is good; the waveform you have is clearly made up up of the basic 60hz and the harmonics at 120 and 180. Any electronics engineer with a knowledge of audio would recognise its typical shape immediately.
Your idea of generating an equal and opposite signal for each of the the three frequencies (all THREE because even if you cannot hear the 60hz yourself, others can... and you cannot quantify what effect it may have on other frequencies later in the chain if there is any non-linearity in the amplifier).
The basic defect in your concept is that you have to generate signals in exact anti-phase and exact opposite magnitude and combine them exactly. To do this you will need variable phase shifters and variable amplitude controls (3 of each), as well as two stable generators of the exact harmonics. All possible; then you set it all up using a scope to see the result, and BINGO!

And then , for some obscure reason, the hum signal coming from the projector decides to change its composition... amplitudes of the components, phase differences between them, etc.. and you have to set it up all over again..until the next time. And, of course, any hum in the recording itself is not taken into account!

Active phase/amplitude, circuits like you suggest work well on STABLE sources, but to eliminate(or greatly reduce) sources that may vary, non-phase conscious FILTERS are the only dependable method. They should be sharply tuned (Notch) filters of high attenuation, so as not to affect other frequencies. They should be placed as early as possible in the amplifying chain, with due regard to their required input and output matching (placed immediately after the playback head would dampen the head impedance and ruin the sound), and where any internally generated noise is much less than the signal.
It is VERY difficult to design sharp filters with good attenuation with just passive components; the active "op-amp" type filter is much better.
Having said all that, and if it makes you feel any happier, the good news is that such a filter actually works very much like you suggest, by feeding back accurate anti-phase signals time and time again to cancel the incoming ones. There will always be some residual at the output, or it wouldn't have anything to feed back! BUT the attenuations can be enormous.
If you incorporate 3 tuned filters(slightly variable for frequency) in each audio channel and adjust them with a scope, the hum will to all intents and purposes disappear, and so will any hum present in the actual recording.
EDIT; Of course, Rick is correct; dispose of as much hum from power and other sources as you can first.
Martin.

Posted by Steve Klare (Member # 12) on November 13, 2009, 01:21 PM:

I have gone the cleanup route. I have been able to replace my filter electrolytics with even higher values in the same packages since capacitance density has improved so much over the decades.

I'm not so sure it is a power supply ripple issue anyway. Since the power supply starts out by full wave rectifying the AC I would expect the fundamental to be 120 Hz and not 60.

I've always had a nagging feeling that a lot of this is pickup due to that big power transformer and electric motor operating in the same space with that sensitive magnetic soundhead. One of the things I do to see if my head is picking up is move my wedding ring near it and see if the distortion of the magnetic field changes the hum (...at last!, a good reason for wearing one!)

If they designed a machine these days it would have a switching power supply operating well into the supersonic frequency range, the motor would be DC and controlled by another high frequency switcher and Hum would be amost nothing! (Just the little bit that sneaks past the PS controller circuit.)

By the way, I'm not that militant about solving this, It's just fun to have a problem to fix without a deadline.

(-beats cleaning the gutters!)

Posted by Martin Jones (Member # 1163) on November 13, 2009, 01:37 PM:

Just an addition following your post Steve. The 60hz comes from general radiation from the mains frequency components in the system being picked up by sensitive devices, such as the heads. The 120 hz, as you say, comes from radiation or conduction from full wave rectified power lines. The interesting one is the 180 hz, because there is no obvious source of that frequency in the machine, so it can only arise from some non-linear amplification of the two other components, which will give rise to sum and difference frequencies (60+120=180, and 120-60=60)further intermingles to produce the final shape waveform)
So replace electrolytics where you can, but don't boost the values too much, so as not to increase ripple currents. And it's the relatively small decouplers on power lines than may have the most effect.
I take it you have a schematic and layouts; if not I can help there.

Martin

Posted by Steve Klare (Member # 12) on November 13, 2009, 01:48 PM:

That's one of my problems here: I've never had schematics for the Elmo.

When I work on a circuit, that's the first thing I expect to have. I feel like I'm flying blind!

Do you have Elmo schematics?

Posted by Paul Adsett (Member # 25) on November 13, 2009, 01:52 PM:

Steve,
Have you tried adjusting the positions of the hum bucking coils? This can make a huge difference in the amplitude of the hum. There is one hum bucking coil beneath the head on the Eumig's and two HB coils next to the head on the Elmo GS.

Posted by Steve Klare (Member # 12) on November 13, 2009, 02:07 PM:

Yes,

I have tried them and with some success. Actually, under regular conditions the machines (especially the Elmo) aren't really "hummy", it's when I interface them to the stereo it gets worse.

I thought that I may have a problem with a ground loop, but even if I hook in a large external (floating) speaker the hum is more pronounced anyway. I think it's a matter of the enhanced bass response bringing out the hum.

By the way, this is mostly an issue in quiet soundtracks and when the machine stands idle. Given any volume from the track, the hum is overwhelmed.

Posted by Martin Jones (Member # 1163) on November 13, 2009, 02:16 PM:

I have the schematics for the ST600, but I see on looking at your post that yours is an ST800. Checking other posts I see that Kevin cannot help, but Super8Wiki shows what appears to be an Elmo Flyer which describes it a s "the popular version of the ST1200" with 800ft capacity instead of 1200ft.
Is it just possible that they have the same circuitry?

Martin

UBB.classic^TM 6.3.1.2