Synchronising two machines

jpmorris

Tape Wolf
The MMC100 unit which I have does not appear to be able to make the TSR-8 chase a MIDI timecode, although it can control the transport at a basic level. This irritates me, although since it cost 2p on ebay, it's not a total disaster.

The MTS1000 is like gold dust, likewise the appallingly expensive ES50.

So. What machines are capable of interfacing with the TSR-8 to lock two together?
Apparently the MicroLynx can do it via the parallel cable. Is this a relatively standard interface? The ML can also talk to the A80, MCI machines and god-knows-what else.. does that mean an 1978 Maglink (War of the Worlds) synchronizer would do the job too? (Although to be honest, I've seen more ES50s on ebay than those...)

Any suggestions?
 
first of all I'm no expert here and I may be somewhat incorrect, but in general as I understand there's no such thing as tape machine "chasing" midi time code. The way it works (if you want your midi device to act as control-"center" in the system): Your midi device sends MMC 'message' (midi machine control) to a device that can 'convert' MMC to start/stop/rew/ff action. As the tape (that is already has SMPTE code on it) moves, SMPTE code is converted to MTC (midi time code) which is being sent 'back' to midi device, midi device chases it and stays in synch with the tape. So the tape is a synchronizing master, but the midi device controls transport.
ok, where am I going wrong? :o
where is BRDTS when we need one? :D pm him, he's the expert.
so, maybe your unit can do what you want? Also what "Synchronising two machines" has to do with MTC or MMC to begin with? ;) What exactly you situation and what are you trying to accomplish?


/respects
 
Dr ZEE said:
first of all I'm no expert here and I may be somewhat incorrect, but in general as I understand there's no such thing as tape machine "chasing" midi time code. The way it works (if you want your midi device to act as control-"center" in the system): Your midi device sends MMC 'message' (midi machine control) to a device that can 'convert' MMC to start/stop/rew/ff action. As the tape (that is already has SMPTE code on it) moves, SMPTE code is converted to MTC (midi time code) which is being sent 'back' to midi device, midi device chases it and stays in synch with the tape. So the tape is a synchronizing master, but the midi device controls transport.
ok, where am I going wrong? :o
where is BRDTS when we need one? :D pm him, he's the expert.
so, maybe your unit can do what you want? Also what "Synchronising two machines" has to do with MTC or MMC to begin with? ;) What exactly you situation and what are you trying to accomplish?
/respects
Let me explain more thoroughly. The current setup is one TSR-8 with a Phil Rees TS-1, which is controlling a sequencer via MTC. Since the sequencer is perfectly adapting to fluctuations in the tape speed even over a 26-minute track, and is also able to do chase-lock if I wind the tape on, there is enough data in the MIDI stream for a sufficiently intelligent unit to slave a second tape machine to the first, and pass the MIDI data on to the sequencer so that IT is slaved too.

In other words, the timecode on UNIT 1 is acting as master, and both UNIT 2 and the computer are slaves to UNIT 1. This was the plan.

If no-one has made a machine which will make a TSR-8 chase incoming MTC data, then UNIT 2 will have to be slaved directly to the timecode off UNIT 1 somehow. This will require a different box. Any recommendations?
 
SMPTE -synchronize the tape machines (so you need a device that can do it). Take SMPTE out of either one - convert to MTC (so you need a device that can do it, which you have) - send MIC to your midi device (which you are already doing). One of your machines will be the master and transport controler. If your synch device/box/unit can do both (synching two tape recorders via SMPTE and also convert SMPTE to MTC and provides midi data out, then that's the only unit you need (?)).
If I only had those boxer , tsr-8s to play with, then I may know a thing or two about how to actually put it all together. I'm just ranting based on general info. :o
 
Dr ZEE said:
SMPTE -synchronize the tape machines (so you need a device that can do it). Take SMPTE out of either one - convert to MTC (so you need a device that can do it, which you have) - send MIC to your midi device (which you are already doing). One of your machines will be the master and transport controler. If your synch device/box/unit can do both (synching two tape recorders via SMPTE and also convert SMPTE to MTC and provides midi data out, then that's the only unit you need (?)).
If I only had those boxer , tsr-8s to play with, then I may know a thing or two about how to actually put it all together. I'm just ranting based on general info. :o

The TS-1 can only convert SMPTE to MIDI and vice-versa. The MMC-100 can control the Tascam transport, but doesn't seem able (or willing?) to lock the transport to an incoming MTC signal, which is a bummer. The TSR-8 won't allow me to select 'EXT' as the speed control, so I'm guessing that it will only do this if it's given a true parallel-port transport control. I'll probably get the TimeLine MicroLynx if I can find one for less than the recorders.
 
I can't tell from your posts whether you have one Tsr8 or two Tsr8's. But, I'll assume for a moment that you have two of them and have explained below what you have successfully done with ONE of them while the other one sits over in the corner waiting it's turn to somehow get integrated into the system-
--------------------------
your quote....."Let me explain more thoroughly. The current setup is one TSR-8 with a Phil Rees TS-1, which is controlling a sequencer via MTC. Since the sequencer is perfectly adapting to fluctuations in the tape speed even over a 26-minute track, and is also able to do chase-lock if I wind the tape on, there is enough data in the MIDI stream for a sufficiently intelligent unit to slave a second tape machine to the first, and pass the MIDI data on to the sequencer so that IT is slaved too.....
---------------------
There is a major typo...or incorrect statement there. What you've correctly stated is ....the Ts1 box will generate smpte signal to slap onto a track of a tape recorder. You record that signal onto a track of a TSR8. The tsr8 then plays...as a master....feeding the smpte to the ts1..which in turn, spits out mtc to a sequencer. The sequencer hums along happily as a slave to the master tape recorder. No problem there. You can do that all day long. A tsr8 can be master for a whole room full of digital sequencers and slaved daws.

But then you stick in the line ....."there is enough data in the MIDI stream for a sufficiently intelligent unit to slave a second tape machine to the first..." Whoa....wrong answer. You can't stick the word "tape recorder" in there. The proper scenario is that there is enough strong healthy midi stream data to drive a room full of slave sequencers and Protools and any other digital animals. Not tape recorders.

Tape recorders can not slave to mtc signals...or pure smpte signals. A digital sequencer can...a tape recorder can not. Unless it's a typo, you're trying to figure out why a great mtc signal that doesn't cause your sequencer to drift can't also reach out there and grab those greasy, wired electric, motors in your tape recorder (perhaps the 2nd phantom TSR8 sitting over in a corner at this point? ) and shove them around in teeny tiny increments to make the tape recorder be a slave....No can do.

---------------------
....."If no-one has made a machine which will make a TSR-8 chase incoming MTC data...."

To verify, no..there is no such thing in the history of the planet of mtc being able to directly make a tape deck slave to a second animal.
-------------------------

.......then UNIT 2 will have to be slaved directly to the timecode off UNIT 1 somehow. This will require a different box. Any recommendations?....

I don't know what UNIT1 and UNIT 2 are. Are you referring to the TS1 and mmc100 as units 1 and 2? Or are Units 1 and 2 maybe two tsr8s?

If the units are two tape recorders....tsr8's...there will be only one way to make them talk with each other in sync. And that is a full blown motor synchronizer that has the ability to read the tach signals off each electric motor (mtc doesn't even know what that means much less can do it) ...and read the smpte signal off EACH machine (mtc can't do that) and then compare all that data AND send out electrical signals to the slave machine to vary it's motors/capstan in teeny tiny increments to keep in frame lock as slave to the master tape recorder...

Bottom line....for two tsr8s (if that's what you have), you have to have a synchronizer. You can not make a motor driven tape recorder slave to another motor driven tape recorder without a full blown synchronizer. You also can not get ONE tape recorder to slave to any black box that simply spits out mtc or fsk or smpte.

An ATS500 or Midiizer are of course the easiest to integrate to two tsr8s because the tach signals and setup are all pretty well hard wired to accept the signals coming down the cables. Plus the dip switches on the back of tsr8's allow settings to match the ats500 or Midiizer.

You CAN get one of the other older mainline synchronizers ...but you WILL then be dealing with making cables and dealing with interfacing Tascam tach signals to match whatever that 3rd party synchronizer wants...which may not be an easy thing to integrate this far down the road from when a lot of that stuff was made.

Of course, once you have a full synchronizer and have the two tape recorders traveling in sync, you can also stick the TS1 anywhere you want in the chain and keep having those digital units slaving all day long to the dual tape recorders....the digital units don't care where they are in the chain as long as they see some mtc to slave to....even if it's from a Y cord from the master tape recorders smpte channel out.
 
Last edited:
BRDTS! U are the MAN! :D
The bottom line is: "Analog machines don't slave!", unless you really really make them ..., they master well though, especially over digital/midi systems , heh heh heh :D
seriously, thanks a bunch for clarification.
If you have time, could you comment on this:
Doin' it on the cheap - Buy a MIDI to SMPTE converter (like Opcode Studio 3). Anything that reads and generates SMPTE timecode, and converts it to MIDI should work. Hook this up to the master 238. Then use an MMC-100 boxes, to give the slave 238 MIDI Machine Control. When you press play on the master, it will play the timecode on track 8, get converted to MIDI by the Studio 3 (or comparable SMPTE box), get sent to the MMC-100 as MIDI, and the MMC-100 will start the slave deck and keep it in sync.
I've read it at this page, and I am not sure if this actually would work as described. :confused:

thanks again,
/respects
 
I forgot all about those old 238's. I bought one new in 1985 or so to use to capture some live music for a project, then sold it soon after. As I remember, it had a pretty good sound considering the amount of stuff being squashed onto the poor little cassette.

I didn't remember that 238s had accessory 2 jacks, but that being the case...and being that two units would involve two motors...it's another instance where a full synchronizer would be needed for sync.

I took a look at the website link with 238 info. It lists all the various Teac/Tascam sync boxes ever made and seems to imply all the interfaces listed would sync two 238's. But in reality, only the ats500 or Midiizer on that list would work.

The article also sort of throws around the term " Midi machine control" along with some of the incompatible Tascam units as if it's a real sync term. In fact, Midi machine control only allows one central transport to activate other transports. But the actual issue of frame or sample accurate sync between multiple units isn't a part of the midi machine protocol.
 
thanX, BRDTS :)
so...? hmmm, I guess it would be more accurate to end that paragraph this way: "...and the MMC-100 will start the slave deck and keep it in sync for a while." :D
/respects
 
BRDTS said:
But then you stick in the line ....."there is enough data in the MIDI stream for a sufficiently intelligent unit to slave a second tape machine to the first..." Whoa....wrong answer. You can't stick the word "tape recorder" in there. The proper scenario is that there is enough strong healthy midi stream data to drive a room full of slave sequencers and Protools and any other digital animals. Not tape recorders.

Tape recorders can not slave to mtc signals...or pure smpte signals. A digital sequencer can...a tape recorder can not. Unless it's a typo, you're trying to figure out why a great mtc signal that doesn't cause your sequencer to drift can't also reach out there and grab those greasy, wired electric, motors in your tape recorder (perhaps the 2nd phantom TSR8 sitting over in a corner at this point? ) and shove them around in teeny tiny increments to make the tape recorder be a slave....No can do.
I respectfully disagree with this assertion. The MTC signal contains an absolute frame of reference which the second recorder can lock against. The MMC100 has, as its input, the time position from the master machine, and the raw SMPTE timecode from track 8 on the second machine. Since it is also connected to the tape transport control on the second (slave) recorder, I do not see any fundamental technical reason why it could not compare the two timecodes and vary the capstan motor speed on the slave to keep it in step with the master (this is after all what the EXT speed controller on the TSR-8 is designed to do).

That is what I was expecting the MMC100 to do, but it looks to me like it only has high-level control over the transport, not down to the capstan motor level. It seems likely to me now that the Accessory-1 parallel interface is necessary to achieve this, and the Accessory-2 serial interface is used to affect high-level control over the tape transport.

Bottom line....for two tsr8s (if that's what you have), you have to have a synchronizer. You can not make a motor driven tape recorder slave to another motor driven tape recorder without a full blown synchronizer. You also can not get ONE tape recorder to slave to any black box that simply spits out mtc or fsk or smpte.
In other words, the MMC-100 is just a controller, not a synchroniser.

An ATS500 or Midiizer are of course the easiest to integrate to two tsr8s because the tach signals and setup are all pretty well hard wired to accept the signals coming down the cables. Plus the dip switches on the back of tsr8's allow settings to match the ats500 or Midiizer.
I was not aware of the ATS500, that's worth investigating. The MicroLynx is what I've half-settled on because I have a schematic for the cable to connect it to a TSR-8.

Any other suggestions for units which are known to be able to interface with the TSR-8 would be appreciated.
 
........"The MTC signal contains an absolute frame of reference which the second recorder can lock against....."

Tape recorders don't slave to mtc or to smpte alone. A synchronizer uses the smpte feeds from two different machines in order to figure out how to control the motors on the slave machine.

..........."I do not see any fundamental technical reason why it could not compare the two timecodes and vary the capstan motor speed on the slave to keep it in step with the master (this is after all what the EXT speed controller on the TSR-8 is designed to do)...."

The Ts1 and mmc100 contain no circuitry to "vary capstan motors". For sync, a real synchronizer has many ins and outs, but the critical three operating all at the same moment are....two inputs for smpte (operating simultaneously in order to compare the incoming smpte from each tape recorder)...and the critical third component, the umbilical cable betwen synchronizer and slave recorder..carrying all the motor tach, motor positional information and nudge controls so that the synchronizer can control the slave machine (via the accessory 2 jack in the easiest Tascam scenario).

The amber "ext" light on a tsr8 is set to light up when appropriate tach signals are present at the Accessory2 jack on the back of the tsr8. The accessory2 jack is the direct connection to the capstan and motors. While smpte is simultaneously shot in and out of a recorder via the rca jacks on a channel, there is no smpte or timing information going through the cable to the accessory2 jack. Just motor commands.

The Ts1 and mmc100 and mts30 and other similar boxes won't work for your tsr8 (as a slave) because they have no internal circuitry for controlling electric motors on tape recorders. Which is actually some pretty complicated circuitry. Which is also why the Ts1 and mmc100 boxes don't cost $2000.
 
BRDTS said:
........"The MTC signal contains an absolute frame of reference which the second recorder can lock against....."

Tape recorders don't slave to mtc or to smpte alone. A synchronizer uses the smpte feeds from two different machines in order to figure out how to control the motors on the slave machine.
Agreed.

The Ts1 and mmc100 contain no circuitry to "vary capstan motors". For sync, a real synchronizer has many ins and outs, but the critical three operating all at the same moment are....two inputs for smpte (operating simultaneously in order to compare the incoming smpte from each tape recorder)...and the critical third component, the umbilical cable betwen synchronizer and slave recorder..carrying all the motor tach, motor positional information and nudge controls so that the synchronizer can control the slave machine (via the accessory 2 jack in the easiest Tascam scenario).
Indeed. It is quite obvious that the TS-1 (or indeed the MTS30) can't do this since its sole connection to the recorder is a signal input, but the MMC100's inability to do so via the Accessory-2 interface was a disappointment. I had assumed that anything capable of driving the transport and doing timecode seeks would also be able to interface to the lower-level functionality and vary the capstan motor speed to balance the two incoming timecodes.

The Ts1 and mmc100 and mts30 and other similar boxes won't work for your tsr8 (as a slave) because they have no internal circuitry for controlling electric motors on tape recorders. Which is actually some pretty complicated circuitry. Which is also why the Ts1 and mmc100 boxes don't cost $2000.
I have no idea what its list price was when new..

Anyway. Just to make sure I've got this all lined up in my head, would a true synchroniser require a control-level connection to both machines, or just the slave? In other words, would it be able to lock against the master from the stripe (as audio) by comparing it against the slave's stripe signal? Or does it need the tach signal from both units?
 
Using an ats500 or Midiizer (mts1000) as an example, you have 2 main choices for hookup if the scenario is two Tascam machines-

Option 1. Smpte output from each recorder's track 8 rca outs...routed to the smpte master and smpte slave inputs on synchronizer. Then you also add an accessory2 cable from EACH machine. One plugs in the synchronizer's acc2-master connector, the other hooks to the acc2-slave connector.

Option 2. Same as option 1 but you only have an acc2 cable hooked between the synchronizer and slave...saving you the time of building/buying a second acc2 cable.

Option 1 is the preferred because option 2 causes a lot of frustrating wait time. Here's why-

In option 1, both machines and synchronizer are talking to each other every moment. Everybody knows what everybody else is doing.

In option 2, there is a missing acc2 cable from the master. (slave acc2 is required). What this means is that during play and record, the master is sending smpte to the synchronizer, the synchronizer throws the slave into play or record....monitors the slave smpte compared to the master smpte, and continuously sends motor commands up the acc2 cable to make the slave motors and capstan stay in frame rate sync at 30ndf accuracy at all times. Sync is tight and works fine between the 2 machines....BUT.....

With the missing second acc2 cable, something happens that's not very desireable in my book-

Without the second acc2 cable, you stop the master tape, intending for example to fast-forward or rewind a minute or so. You throw the master into rewind. Guess what? The slave just sits there during the master's fast forwarding and rewind. Why? Because during ff and rew, the master is not sending smpte signal out the heads...as you know, head signal is muted at ff and rew ...otherwise you'd hear loud whining of the tape with the resulting sound frequencies breaking your windows, irritating your dog, blowing out your speakers....etc etc. Besides, during ff and rew, smpte is breezing by so fast that nothing could read it anyway.

In a sync setup between 2 machines....any two machines, when a master is put into ff or rew, the way the slave machine gets a command to chase a master and ALSO rew or ff is via motor tach signals. A synchronizer (the intelligent being it is) , sees that "hey the master just went into ff or rew, so I'm gonna stop looking for smpte and instead I'm gonna start counting the motor tach pulses to see if the motor is going forward or backward....and then I'm gonna tell the slave motors to do the same thing. Synchronizer looks at the master the instant it starts ff or rew, figures whats going on in a nanosecond, shoots the command to the slave to do the same thing....both machines zoom off. As soon as you hit stop on the master when you get to whatever point on the tape you want, the synchronizer immediately sees what you did, takes a nanosecond look via tach as to approximately where-in-time you stopped, shoots that command to the slave and instructs the slave to stop in the same time neighborhood. A nanosecond later, slave is there and stops. Th instant you hit "play" on the master, the synchronizer sees the tach speed...sees the smpte, knows you went to "play" and shoots the message to the slave to start playing....then instantly looks at the master smpte and slave smpte to see how far they are off from each other...shoots the commands to move the slave motor faster or slower to get to the exact point as the master, then falls into "lock". This is called chaselock.

Without the second acc2 cable, when you throw the master into fast-forward, the synchronizer sees no motor tach signals...because there's no acc2 cable from the master. All the synchronizer knows is that there's no longer any smpte from the master. So it shoots a message to the slave saying "stop ..because I don't see the master and I don't know what's going on yet so let's just stop and wait here".

If you had been fast forwarding the master recorder a minute forward on the tape....which takes a few seconds as you know....the slave sits there. When you arrive at the tape destination you want and then hit "play", the synchronizer sees the new smpte and says, "whoa, the master's now 3 minutes from here." The synchronizer shoots the command to the slave to now ff or rew to where the master is.......which you have to sit there and wait for. Or....just keep playing the master and the slave will eventually catch up and lock in a few more seconds......which isn't ideal if you're doing something like say, fast punchins.

Without the second acc2 cable, bottom line is you spend quite a bit of time waiting for TWO tape recorders to ff and rew at TWO different times. With two acc2 cables, continuous contact is always established and the two machines act more or less as one...even during fast forward and rewind.

In the case of Tascam.

Now...let's say you have a Studer a824 and a Midiizer and a TSR8 that you want to synchronize to the Studer. A strange setup if you think about it, but hey...it could happen. In this instance, the Studer is using whatever motor tach signals it uses and those may or may not be appropriate to the tascam synchronizer. You could build some sort of interface from the Studer, but in reality, the easiest thing to do where ONE of the recorders is a non-Tascam is to simply operate with the Option 2 scenario and deal with the lag times during ff and rewind between the two machines.

If you have two tsr8's....a synchronizer...and two acc2 cables (built yourself or purchase via a couple of places that make them), you're pretty well set for full sync.
 
BRDTS said:
In option 2, there is a missing acc2 cable from the master. (slave acc2 is required). What this means is that during play and record, the master is sending smpte to the synchronizer, the synchronizer throws the slave into play or record....monitors the slave smpte compared to the master smpte, and continuously sends motor commands up the acc2 cable to make the slave motors and capstan stay in frame rate sync at 30ndf accuracy at all times. Sync is tight and works fine between the 2 machines....BUT.....

[the machines lose sync when the master does a seek]
That is something I honestly hadn't considered, although for my purposes I could probably live with it. Mostly it's going to be laying down sequencer-controlled synthesizers, which wouldn't usually need a lockup anyway. For vocals I could get away with a reduction mix on the second machine as a guidepost.
The point where I'll absolutely need sync is going to be the final mixdown - pressing RTZ on both machines is probably not going to be too much of a burden. For video work or suchlike it would be a different story.

Anyway, that has been quite helpful. Thanks.
 
Thanks very much BRDTS for that explanation. I've just got a couple of 688s, a couple of 238s and an ATS-500 which I'm going to try syncing up (not altogether of course, but a pair at a time).

I now understand what the ATS-500's role is and what I need to connect to achieve this.

Al
 
The topic of this thread reminds me of a record I have by the late Hollywood studio trombonist, Lloyd Ulyate:

LLOYD ULYATE & HIS TROMBONE
1.The Trolly Song
2.In My Solitude
3.Anything Goes
4.I Loves You Porgy
5.Holiday For Trombones
6.Gone With The Wind
7.The Continental
8.Trombolero
9.Steppin' Out With Baby

Unfortunately, this album is LONG out of print (recorded in the early 60's); however, it was engineered by the legendary Dayton "Bones" Howe. It's a fabulous album.

Lloyd recorded all the trombone parts and Bones Howe had to precision sync two Ampex (300's ?) to eliminate a harmonic distortion problem that cropped up due to stacking so many parts. The whole story is on the liner notes. If I can get permission to reproduce them here, I will. Otherwise, I'll try to find a web link that I can post with more info. This album was considered a milestone in multi-track recording at the time - long before it became a common practice.

MCJ
 
Fwiw, a Sony APR24 will sync to smpte with nothing else required. Just enter the offset and off she goes.

Earliest syncing was accomplished with a 60 cycle resolver which provided phase lock. But there was no address info. The engineer marked the start postion on the tapes at the center of the head with a china marker and hit play on both machines at the same instant. And hope for the best.

Problem was is that it could lock/resolve and play at the same tempo anywhere along the tone but both machines may or may not be at the same point in the song, ie: the resolver could synch the tape speeds but not program. Consequently, both machines were rarely synced together except at mixdown. Iow, a rough mix bounce was done on the second machine while locked and that tape was used alone until mix time.
 
The MMC100 unit which I have does not appear to be able to make the TSR-8 chase a MIDI timecode, although it can control the transport at a basic level. This irritates me, although since it cost 2p on ebay, it's not a total disaster.
Hey JP! Curious if you still have this MMC-100 unit laying around? I know this post was from like 15 years ago (so it’s unlikely) but I’m desperately seeking one of these to buy for my own setup.
 
Hey JP! Curious if you still have this MMC-100 unit laying around? I know this post was from like 15 years ago (so it’s unlikely) but I’m desperately seeking one of these to buy for my own setup.
Incredibly, I still do. The unit and the manual, but no power supply for it. Shipping to the US is likely to be expensive but doable.
 
Back
Top