Major 1979 Allen MOS1 Tuning issue

[zylstra555]

We had our 1979 Allen MOS1 delivered last Wednesday, and found out that the organ is quite out of tune. Also, the principals seem to not be working unless the main or flute vibratos are on, as soon as you turn those off they die down to an inaudible level again.
The organ is about 2 half-steps sharp plus about 10 cents -- we are currently using the transposer to help fix this issue.

How can we first fix the tuning issue, then the issue with the principals?

It'd be nice if I could get this going again by Sunday (tomorrow!) for the first day of Advent.

EDIT: The main vibrato tremulant alone will cause the principals to come through.
When we got the organ, the principals were working, but would cut off after awhile. If you turned the organ off then on again, they would begin to work again. I am not sure if that helps, but I figured I ought to mention it.

[zylstra555]

Okay, we came up with a temporary fix for the service tomorrow:
We moved the transposer down 2 half steps, and tuned the master tuning thingymajigger in the back of the organ (the frequency divider?). Still hoping to get the rest of it back to where it needs to be if possible.

[myorgan]

Okay, we came up with a temporary fix for the service tomorrow:
We moved the transposer down 2 half steps, and tuned the master tuning thingymajigger in the back of the organ (the frequency divider?). Still hoping to get the rest of it back to where it needs to be if possible.

Could it be that the master tuning is raised 2-half steps sharp? I didn't think that was possible, but who knows? I certainly hope everything turned out OK today!

Michael

P.S. We did Thanksgiving music today because the outgoing interim decided to do Communion last week instead! Grrrrr.

[zylstra555]

I can try messing with it again, but when I was adjusting the master oscillator (I should note that I assume that that is what you would call it, I took a picture that I should share) it went to a certain point in tuning and wrapped around to where it started, like it can only take it so far and it just goes back the beginning.
More about half the stops not working:
Today I was playing extremely loud while practicing "for fun" with all of the pedal stuff down, and suddenly for about one minute all of the stops were working again -- it was quite loud and quite wonderful. This tells me that something could be loose, but I've checked everything I know to check -- what else am I missing? Either something is loose, or some weird fluke happened during that time.

[myorgan]

Today I was playing extremely loud while practicing "for fun" with all of the pedal stuff down, and suddenly for about one minute all of the stops were working again -- it was quite loud and quite wonderful. This tells me that something could be loose, but I've checked everything I know to check -- what else am I missing? Either something is loose, or some weird fluke happened during that time.

You might be onto something! I know this won't help, but hearkening back to a senior member on this Forum, JBird604 (I miss him here), he would recommend re-seating the boards. Quite possibly a connection is loose and just needs to be re-seated.

Let me do a search on one of his posts about that, and link it on this thread. If I remember right, he was referring to ADC organs, but the advice may apply to MOS-I & MOS-II organs as well. I'll look more when I have time.

Michael

[don60]

The dividing point for the main and flute channels is the digital outputs of the main board (the digital inputs of the digital/analog converters). If the problem is occurring before this point, the only possible remedy in the field is to try reseating the socketed MOS chips. If the problem is occurring after this point, interchanging the main and flute inputs to the DAC board should make the flute stops disappear instead of the mains.

Turning on the tremolo causes the tone generators to read from different entries of the stored frequency table than they do with non-tremmed tones. Your observation that using tremolo restores the missing stops, together with your observation that the entire organ is significantly out of tune, leads me to believe that you have a problem with the frequency generator portion of the main board.

If you feel confident doing so, try reseating the MOS chips.

Don

[myorgan]

Either something is loose, or some weird fluke happened during that time.

This is the thread to which I was referring: http://www.organforum.com/forums/sho...ight=petroleum

[zylstra555]

Thanks for the link Michael.

Don, I thought MOS1 organs used individual hardware components for tremolo, having nothing to do with the stored waveforms. I assume this because there are two tremolo devices in the organ, one for flutes, and one for everything else. Come to think of it, the flute tremulant looks original while the main tremulant looks newer like it was replaced.... hmmm....

This week I will go through and look at all of the audio connections again, organ on and a key pressed down with the trouble stops. If I fail to find anything, I'll reseat the boards. I am slightly reluctant to reseat the boards since the edge connectors are also screwed in, but I think that's laziness speaking. I have a degree in Network Administration, but I would be reluctant to reseat the chips except as an absolute last option.

I have seen the DAC, but I do not understand its function. Can someone explain this to me?

(...I had to spell out tremolo three times before I got it right...)

Thanks again for all the help. Hopefully we can figure this out!
Best,
Jesse Zylstra

[davidecasteel]

Technically, in music, tremolo is a periodic fluctuation in intensity (volume) and vibrato is a periodic fluctuation in pitch. Pipe organs have a function called Tremulant which causes the air pressure fed to the pipe feet to fluctuate slightly at a fairly slow rate (5-7 pps); this causes mainly a fluctuation in intensity, but also some variation in pitch. Most early electronic organs had a Tremolo function, and it generally only varied the intensity; some higher-end instruments had a Vibrato function instead--many of those used a phase-shifting circuit to essentially create a low-frequency FM modulation of the notes played (which is what vibrato actually is).

If a digital instrument is doing a tremolo, I would not expect there to be any different values in an eProm somewhere involved; OTOH, if a vibrato (or pipe-like combination) is desired, then different values to produce different pitches possibly would be required.

In other words, if a fluctuation in intensity only is needed, that could be imposed after the DAC has converted the binary values to the analog sound signals, but a pitch alteration would probably need to access a different set of binary values. (Alternatively, it would be possible to affect the pitches of all notes played in a channel by just varying the frequency of the master clock at the proper rate; and the intensities of the notes played could also be varied by switching in different sets of binary values with less intensity. I think the latter would be more complex than necessary, though.)

Z, a DAC is a Digital-to-Analog Converter and its function is to take as input the binary values that represent the time-slice amplitudes of the audio signal waveform and produce a varying-voltage analog audio signal out of them. This can then be amplified and sent to the speakers to produce the sounds that you hear. The simplest form of a DAC is just a large network of resistors that can be quickly switched into or out of a voltage divider to produce changes in the voltage output.

David

[radagast]

Technically, in music, tremolo is a periodic fluctuation in intensity (volume) and vibrato is a periodic fluctuation in pitch. Pipe organs have a function called Tremulant which causes the air pressure fed to the pipe feet to fluctuate slightly at a fairly slow rate (5-7 pps); this causes mainly a fluctuation in intensity, but also some variation in pitch. Most early electronic organs had a Tremolo function, and it generally only varied the intensity; some higher-end instruments had a Vibrato function instead--many of those used a phase-shifting circuit to essentially create a low-frequency FM modulation of the notes played (which is what vibrato actually is).

If a digital instrument is doing a tremolo, I would not expect there to be any different values in an eProm somewhere involved; OTOH, if a vibrato (or pipe-like combination) is desired, then different values to produce different pitches possibly would be required.

In other words, if a fluctuation in intensity only is needed, that could be imposed after the DAC has converted the binary values to the analog sound signals, but a pitch alteration would probably need to access a different set of binary values. (Alternatively, it would be possible to affect the pitches of all notes played in a channel by just varying the frequency of the master clock at the proper rate; and the intensities of the notes played could also be varied by switching in different sets of binary values with less intensity. I think the latter would be more complex than necessary, though.)

Z, a DAC is a Digital-to-Analog Converter and its function is to take as input the binary values that represent the time-slice amplitudes of the audio signal waveform and produce a varying-voltage analog audio signal out of them. This can then be amplified and sent to the speakers to produce the sounds that you hear. The simplest form of a DAC is just a large network of resistors that can be quickly switched into or out of a voltage divider to produce changes in the voltage output.

David

David,
Would you say that a DAC changes an AC electrical signal to a DC signal?

[don60]

David,

Your definitions of tremolo and vibrato agree with mine. Allen has a different view, however. Their digital tremulants were implemented using frequency shifting, at least in the MOS 1 through ADC instruments. As evidence, I quote from one of several discussions in the MOS 1 service manual regarding frequency generation:

"There is a separate Speech Articulation Control for each manual, but none for the Pedal. The Speech Articulation circuit provides a pipe-like pitch shift upon the attack of any key. This is accomplished by temporarily gating the Tremulant control circuit for that manual immediately upon the attack of any key. The Random Motion Circuit does somewhat the same thing--that is, it turns the Tremulant Control circuits for the respective manuals on and off--but in a random pattern. This imparts a desirable random activity to the sound. With both of these functions--Speech Articulation and Random Motion--the Tremulant oscillator signal from the Keyboard Array to the MOS board is locked off. However, when the actual Tremulant stops are used, the Tremulant oscillator signal to the MOS board needs to be reactivated. This process of turning the Tremulant oscillator signal on and off is the function of the Tremulant Gate board to which the Tremulant Stops are wired."

Elsewhere, the manual mentions that the digital celeste (as opposed to the analog celeste rank used in some early models) is implemented by stopping the tremulant oscillator with tremulant enabled, which causes a constant sharp frequency shift in the affected division.

I know from my own circuit tracing that tremulant in the ADCs is implemented in a similar way--a low-frequency oscillator steps the frequency generator output through a succession of values in the vicinity of the nominal note frequency.

The Allen TREM III generators were outboard analog modules used to add what they called Vibrato to certain divisions. Judging from the adjustments available on the front panel, they imparted both an AM and an FM effect.

To be fair to Allen, theory tells us that narrow-band frequency modulation and amplitude modulation have identical frequency spectra (ignoring phasing), so what would properly be called tremolo ought to sound quite similar to what would properly be called vibrato, at least at modest levels of intensity.

Z, I believe that the tremolo board you mention is just the oscillator driving the main MOS board's digital tremulant logic. It is not an analog amplitude or frequency shifter. I stand by my diagnosis that you appear to have a frequency generator problem. A disabled trem oscillator could produce a slightly sharp tuning, but not by two or more tones. I think you have a MOS logic problem somewhere on the main board.

Don

[davidecasteel]

David,
Would you say that a DAC changes an AC electrical signal to a DC signal?

No, not at all. Basically, a DAC just accepts a binary array of 1s and 0s as input and generates a DC voltage equivalent to the binary value entered. If only one value is input and it never changes, then a constant DC voltage would be output.

In digital organs, DACs are used to generate the audio signals corresponding to the notes being keyed. The different voices (stops) are stored in the instrument as sets of binary values that represent the amplitudes of the audio signal at many points in a cycle, and (in simple terms) these sets of binary values are presented repeatedly over and over to the DAC at a rate such that they are repeated at the keyed pitch. As each new binary value is presented to the DAC, its output voltage changes to match and the result is a voltage that varies up and down to replicate both the waveform (responsible for the timbre) and the pitch. The DAC receives a binary value as an array of 1s and 0s and gives a matching DC voltage; other circuits decide what values to send to the DAC and how fast, and the result is an audio signal.

As a simple example, let's say I have a 4-bit DAC. If I send a binary value of 0101 to it, it will immediately produce an output voltage proportional to the number 5 (which is what 0101 binary equals in decimal); if I then give it a value of 1101, the output will be proportional to decimal 13. If I feed different binary values into the DAC at a fast enough rate, a signal will be generated that approximates an audio signal (some filtering is usually necessary to smooth out the signal, though). (I am assuming a DAC that does not have a "sign" bit--which is how Don has indicated the Allen DACs work; the Allen DACs use an 8-bit binary word, though.)

To reiterate: a DAC just converts a set of binary 1s and 0s into a DC voltage proportional to the binary value. The binary data is "Digital" and the DC voltage is "Analog", hence "Digital-to-Analog Converter (DAC)".

David

[circa1949]

Just a minor clarification. Of course David is correct that if the digital signal stayed stuck on one value, the DAC would output a single DC voltage. (Although my understanding is that there are "current" and "voltage" DAC chips that use different types of output buffers to generate a normal AC audio signal, so presumably some would instead output a constant current value...not that it really matters to the big picture given that P=I^2R)

However, that wouldn't be audio! All normal analog audio signals are alternating currents with respect to 0V, cf: http://en.wikipedia.org/wiki/Line_level Again, not that David didn't know this, just that it could seem unclear to some readers.

[circa1949]

I think you have a MOS logic problem somewhere on the main board.


This raises an interesting question about MOS organs. Clearly, these organs constitute some of the oldest operating digital circuitry IN THE WORLD. I know a fair amount about "government information technology". I've been in computer rooms that were about the size of 1/2 a football field that used to be filled with mainframes, whose functions were mostly consolidated into one single machine the size of a large refrigerator. (with the rest of the space taken up by the client/server model of computing) Maybe somewhere there are some old ICBM using such chips, but I doubt it. What is the expected lifetime of them before the natural decay of the universe just takes its toll? One does hear of MOS organs needing new boards, but also that some are still ticking along just fine. And one doesn't know whether they died because of something like a lightening strike. At any rate, I'm sure they have been around much longer than the manufacturer's expected life time. (i.e., the manufacturer of the chips, not Allen themselves) Even though Allen will service them for a price, it would be foolish to pay a lot for even a large MOS organ if you expect to primarily use its tone generation components. Of course for conversion to Hauptwerk, they are probably ideal in some ways. Well built, with wiring and keying systems that are standardized enough that the market has produced a MIDI kit for both MOS-1 and MOS-2.

Don't think I'm implying modern circuitry will last longer, it almost certainly won't. Although the problem of "tin whiskers" with "ROHS" lead-free solder have supposedly been addressed with new formulations, I wouldn't bet my life on it. For one thing on an SMD board, the whisker only has to grow about 0.3mm, versus maybe 2mm on a through hole/DIP board of yore.

[davidecasteel]

(There were 3 posts here--2 by me--that should have been posted on the Allen Tone Cards thread in this forum. They have been replicated there now and removed from here. Sorry about that.)

David