My TBC-1000 distribution amplifier bypass mod!

[aramkolt]

My take on the TBC-1000 distribution amp bypass mod

There are a few old threads on this idea, but thought I'd throw my design out there for some criticism/comments. It's not quite fully wired up yet but basically just needs the power connector and the front switches connected.

The three switches on the front will be used for:

1. Power
2. Fan
3. Composite/S-Video Select

General observations:

The distribution amp on my version only has one diode and one extra 470uf cap which will be easy to add in the signal path and I may just add it at the power jack itself. Alternatively, the Diode could be put on at the input side of the mini power board and a larger electrolytic cap could be used there which would accomplish the same thing.

The fan may add noise to the video signal, but that remains to be tested. Alternatively, the fan could be powered from a separate power supply and that should not produce any noise. Yes, the fan is almost certainly overkill and not "necessary", but for a device this expensive, seems reasonable to keep it cool to prolong it's life and/or be able to use it 24/7 without heat issues as long as it doesn't worsen the output quality.

The mini power board was moved away from the video inputs because I figure that could cause some RF noise and it kind of gets in the way of wiring over there anyway. I originally thought the mini power board was a linear regulator that dumps the extra voltage as heat for a very low-noise output, but it turns out it's a DC to DC switching IC which may be no better than hooking it up directly to a 5V switching power supply. Seems it's really only there because the distribution amp requires more than 5V which we are no longer using.

The video input cables are purposely NOT passed over the logic IC's as they may emit some RF noise.

I haven't checked how "noisy" the voltage output is from the mini power board, but if there are noise issues, I plan to wire up a simple LM7805 which is a linear regulator or use an ultra low noise power supply based on the "Studer 900" that you can get on Aliexpress for around $70. Either should be very low noise and if the linear regulator was used, It would be bolted to case as the original distribution amp's 8V regulator was. I also wasn't sure how much current the TBC-100 uses, but turns out the switching IC has a max current output of 0.6A per the data sheet, so the whole thing is only using 3W max apparently.

Based on the above, it's likely you could delete the mini power board altogether and just feed 5V to the TBC-100. If you want that to be a relatively low noise source, you could probably even do it with regular USB power from a computer or good quality USB charger.

If eventually converted to a 5V only system, I would hesitate to use anything other than a USB plug as a 12V could get plugged in by mistake as a lot of other equipment is 12V. That, and the TBC1000 is also labeled 12V on the outside.

I purposely left a gap in the front for airflow from the fan which is directed over the path of chips that usually get warm. Could also end up swapping that fan for a squirrel cage blower fan that would probably get more air under the board as it is slightly raised for top/bottom cooling.

It would be just as easy, or possibly easier to use BNCs for S-Video and Composite output. My logic with the direct S-Video cable is that those are of decent quality and makes for one less connection. Also didn't seem like there were a lot of good options for S-Video jacks available, but I'm sure I could 3D print some sort of holder for those jacks as well. I do have high quality male S-Video connectors that I could run two of those Belden mini RGB coax into (that's what the red and green coax is that I used on the RCA jacks).

Could directly solder the inputs to the TBC-100, but I think the gold plated connections on the Dupont connectors are good enough and allows for easy cable swaps to other cables if desired.

I do plan to add some clamp-on ferrite beads to all of the video inputs - not sure if it will help with noise, but it shouldn't hurt.

I'll eventually post some results of what the output looks like (both with fan on and off) and if others are interested in having me do this mod or similar to theirs or recap their TBC1000's, I might be open to doing that depending on what other projects I have going on at the time. I'll also post what the thermal camera shows with the fan on vs off to see what sort of temperature differences there are with the fan.

IMG_3729.jpg

You must be logged in to view this content; either login or register for the forum. The attached screen shots, before/after images, photos and graphics are created/posted for the benefit of site members. And you are invited to join our digital media community.

IMG_3730.jpg

You must be logged in to view this content; either login or register for the forum. The attached screen shots, before/after images, photos and graphics are created/posted for the benefit of site members. And you are invited to join our digital media community.

IMG_3732.jpg

[latreche34]

As this was initially designed to fit "passively" inside a desktop PCIe slot (TBC100?), it is equipped with a 4 pin Molex 12V/5V input socket, I believe it is better to use a high quality external power brick or an internal one with better shielding and ripple characteristics than the cheesy one fitted with. Not sure if the missing power regulator is required to be installed to make the connector usable again. Maybe looking at the original TBC100 design can give a clue on how it was used inside a desktop computer.

It maybe even useful to use the original RCA and S-Video sockets by finding the right ones and resoldering them to eliminate any wiring, This will require a different case or reposition the board differently to have access to the video ports.

Your fan design is not ideal, For heat extraction to work the fan has to be blowing out the box and sucks cold air from another vent and no other holes should be present in the case other than the suction and extraction ports, that way fresh cold air is sucked in, warm air is blown out. The way you installed it is called circulation mode, it works good when air is cold but as air warms up it will be just circulating hot air. Keep in mind that cooling an object is by extracting its heat and moving it away, not by just moving the same air over and over.

[lordsmurf]

Whoa, whoa, whoa. Just to add an extremely important note to this thread:
You better be careful ripping a VP-301x card out of a TBC-1000, and cramming it into a computer, thinking it's a TBC-100 card. Or vice versa. There are some very stark differences, and could nuke the card.

[aramkolt]

Yeah, I'm not sure what the noise characteristics are currently, but the card itself just needs 5V. The higher voltage adapter was just needed for the distribution amplifier.

I can't seem to find any actual close up pictures of a TBC-100 to see what IC is used for 12V step-down, but guessing it is an LM7805 which can deliver up to 1.5A which is more than double what the switching DC to DC IC in this does, but that does in turn waste the excess voltage as heat in exchange for low noise, so Ideally you'd want to bolt that to the case.

There's a big square hole next to my switches in the front that the fan is directed at and most of the air comes from slightly behind it. It's not the coldest air source, but the entire thing only draws 3W at most (given the max current of that switching IC), so it's not exactly like trying to cool a CPU. I found that the hottest chip (the Conexant one) ran at about 117 degrees just at idle, after being on for a few minutes - not sure if it gets hotter in actual use. After switching on the fan, it dropped 15 degrees to around 102 which is just above body temperature, so that shouldn't be harming any components anytime soon haha. I actually expect it to get colder than that with the lid on because then the air that escapes upward and out of the case will be directed at the exit hole more efficiently. I'll take a video in a bit to sort of show how much airflow comes out that hole even tough there's no ducting per se.

I actually do have a source for the original style S-Video connectors - Old Apple ADB keyboards - the mounting pattern is the same and they are shielded on the outside of the plastic block as a bonus and they look identical to the "output side" one on the TBC-100. The input side of the TBC-100 used one odd connector that detected whether an S-Video plug was present, but that can be done with any regular switch to go back and forth between S-Video and composite. In my design, that's the middle switch on the front. I would have gone the direct S-Video jack route, but it results in a very tight bend in the cable as there's not much space on that side of the enclosure. Yes, you could drill big holes in the side of the enclosure and give them a straight shot at the plugs, but I didn't feel like doing that at this point anyway as I'm curious if there's any noticeable issues without case mods.

As for the computer ATX power connector, it is odd that they just didn't leave the 12V line disconnected altogether since it just needs the 5V line - perhaps back in the day not all power supplies had 12V and 5V available? Basic testing on mine shows the 5V ATX line and ground has continuity with the tiny plug that is used for power injection on the TBC-1000. The 12V line goes to the pin right next to the 5V input which is exactly what a LM7805 would take in 12V input and have it regulate down to 5V. The way mine is wired, if you plugged a standard ATX supply with an LM7805 in place, it looks like it would draw power from both the 12V and 5V lines while converting the 12V into 5V. With the LM7805 missing, it should only draw from the 5V source and no damage should occur as the 12V doesn't go anywhere apart from a few caps meant to stabilize the incoming voltage. I'd probably cut the 12V pin on the ATX connector just as a precaution anyway if trying the ATX power route.


I didn't look at it too closely, but there's a jumper pin around the power supply area that I am unclear what it would do if removed, but shouldn't be too hard to figure out with some further study.

Attached are the thermal camera photos and I'll try to see if I can rig up some streamers for a short video on the air output from the front with the fan on to give an idea of airflow and how quiet the fan is.

Fan Off.jpg

You must be logged in to view this content; either login or register for the forum. The attached screen shots, before/after images, photos and graphics are created/posted for the benefit of site members. And you are invited to join our digital media community.

[aramkolt]

Here's what it looks like wired up. The foam on top of the fan is to give it some stabilization from the top of the case being pressed down against it. The bottom of the fan is held down with adhesive velcro.

IMG_3738.jpg

You must be logged in to view this content; either login or register for the forum. The attached screen shots, before/after images, photos and graphics are created/posted for the benefit of site members. And you are invited to join our digital media community.

[aramkolt]

Fan noise is kind of hard to gauge here since you can hear without the fan there's a good amount of ambient noise and the phone's microphone is wanting to amplify everything to pick up on the small details - such is the less discussed "audio automatic gain control" haha. It's possible to feel the breeze about 8 inches away and very obvious up close.

[latreche34]

Quote:

Originally Posted by lordsmurf

You better be careful ripping a VP-301x card out of a TBC-1000, and cramming it into a computer, thinking it's a TBC-100 card. Or vice versa. There are some very stark differences, and could nuke the card.

It doesn't, The card has no pins where it connects to the PCIe slot, hence me putting "passively" inside quotes.

[latreche34]

Quote:

Originally Posted by aramkolt

I would have gone the direct S-Video jack route, but it results in a very tight bend in the cable as there's not much space on that side of the enclosure. Yes, you could drill big holes in the side of the enclosure and give them a straight shot at the plugs, but I didn't feel like doing that at this point anyway as I'm curious if there's any noticeable issues without case mods.

The idea is not to drill holes, but rather rotate the board 90 degrees and see if the missing video sockets align with the holes at the back side of the unit, this may require adding new stand-offs to mount the board on the bottom chassis.

Quote:

Originally Posted by aramkolt

As for the computer ATX power connector, it is odd that they just didn't leave the 12V line disconnected altogether since it just needs the 5V line - perhaps back in the day not all power supplies had 12V and 5V available? Basic testing on mine shows the 5V ATX line and ground has continuity with the tiny plug that is used for power injection on the TBC-1000.

What if you disable the 12V pin at the power source in the power brick and just leave 5V, Are you sure the TBC-100 board does not use 12V at all?
Also there is a fan plug next to the 4 pin molex, have you checked if it has any power at all to power up the added fan?
If you can post some high resolution pictures of both sides of the TBC-100 board we maybe able to reverse engineer it and add some suggestions for future mods.

After reviewing some online photos of the board it looks like you could use a high quality non switched 5V power supply and get rid of that power board and the 12V power brick supplied with the TBC-1000.

As to why there is 12V pin on the TBC-100 board, My guess is to drive the fan connector next to the 4pin molex, They figured this card will add heat inside the desktop so an extra 12V fan is needed to extract that heat, so they provided 12V for it, This is my theory, I haven't found any evidence to backup this claim.

[aramkolt]

I did consider board rotation, but the holes in the back where the video connectors originally were are actually also not large enough to fit the injection molded area of an intact S-Video cable through them, so you still end up either drilling holes or making those existing holes a lot larger which I guess is reasonable since it's a mod after all.

You might be onto something with the 12V of the TBC100 being for a separate fan. That would explain the extra larger connector next to the ATX plug that is both connected to the 12V line, though the 12V line can also be regulated down to 5V if that linear regulator is installed (as it always was on standalone TBC100's) - so it is as if it can run on either 12V or 5V, but really only needs the 5V.

It definitely only needs the 5V because that's all that goes into it on the TBC-1000.

I did some basic playing around with it and the composite connectors I used didn't make very good connections, so I've already swapped those for BNC's. The fan running didn't seem to add any extra video noise to the signal which is good, but it wasn't a very scientific test - I'll have to drag out the oscilloscope later. If it was converted to a 5V system only, whatever fan used could be powered by a separate power supply or you could just use a 5V fan.

All things considered, I think the other moders such as DiopterDoctor were onto something by using all BNCs and soldering coax directly to the board - gives a high quality connection and no case modding required and it gives the least distance of unshielded video wiring - other than maybe the S-Video jacks. S-Video likely doesn't make as good of contact as BNC typically would though. There's less chance of card stress by mounting the connectors to the case itself. I'm not particularly concerned about the quality of existing BNC to S-Video cables out there since I would just make my own out of that same Belden RGB mini Coax (the red/green jacketed cable you see on the composite ins).

I do have one of those "Studer 900" power supplies from aliexpress on the way and I'll probably eventually use that, but I will whip out the oscilloscope and just see how the original supply, LM7805 linear regulator, and the Studer 900 compares under load once I have everything here to actually test. Now I'm just second guessing myself if the transformer type used in the Studer construction matters haha. Either way, should be much less noisy than what't there now.

[latreche34]

Quote:

Originally Posted by aramkolt

I did consider board rotation, but the holes in the back where the video connectors originally were are actually also not large enough to fit the injection molded area of an intact S-Video cable through them, so you still end up either drilling holes or making those existing holes a lot larger which I guess is reasonable since it's a mod after all.

My idea is to get the connector out of the holes not the cable through the hole into the connectors, But if the centers of the existing holes do not align with the centers of the sockets on the board it will be hard to enlarge the holes with just a drill bit since the desired holes will be slightly offset from the existing holes.

Yes it's weird having a 5V input and then use a regulator to make another 5V out of 12V, But this is not necessarily wrong, Having two paralleled identical voltages will only increase the amperage and average out the voltage, Maybe the reason of doing that is to cope with load demand. Also it could be that there is a way that if 5V is detected in the 4 pin the regulator will shuts off the extra 5V. But all these requires reverse engineering the board and knowing exactly the schematic. We are just speculating here.

[aramkolt]

Here's the final version I came up with.

I'm using a "Studer 900" linear power supply which is supposed to be known for very low noise power specifically targeted towards DAC use (mainly in audio systems). This is set to output 5V via fixed cable into the TBC1000. The idea is that since this is now a 5V system, I do not want to chance plugging the wrong adapter into this and having a male power connector should easily prevent that possibility. I haven't actually hooked it up to an oscilloscope to see just how low noise it is. If I was buying one again, I'd probably get a variant that has the power on LED in the front instead of the back, but I'm sure functionally the variants are pretty similar. There's a variety of transformers they can use (R-Type, Toroidal, and I believe plastic enclosed toroidal). I got the plastic encased toroidal since it's the most compact and we are using less than well under 1 amp here. The supplies are generally capable of delivering up to 5 amps @5VDC.

I imagine using a simple LM7805 linear regulator inside the case and keeping the 12V input would still be an upgrade from the original switching power supply inside the unit that was used for 12V to 5V conversion in terms of signal noise. That could be heat-sunk to the case in the same place that the distribution amp's original TO-220 chip was.

I removed the Dupont-type headers because they theoretically act as little unshielded antennas, and it forces the cable to have a fairly aggressive bend at the top.

All video wiring has been moved off to the side so that nothing goes over the ICs on the board. I think the issue where they can pick up noise is probably more an issue if the inputs/outputs get anywhere near the processing chips which probably do emit some radio frequency noise.

Cabling is 75ohm MiniRGB Coax with a single solid conductor which I believe is 22awg. You can see it is double shielded with heavy braid on the outside and a foil inner shield. This cable is meant to do runs of several hundred feet, so It's definitely thicker than what you need, but it shouldn't be a source of significant loss anyway.

All of the BNC connectors are electrically insulated from the case where they are mounted as they use 3D printed plastic washers to fill the holes appropriately and this may prevent some form of video "ground loops." [theoretical problem fix, no idea if it actually helps, but shouldn't hurt].

The number of ferrite beads are probably overkill and there likely sin't much in the way of noise way off to the side of the card, but there's room, so figured might as well.

I haven't tested if the fan produces any electrical noise on the 5V power rail, but if it did, the fan can be switched on or off from the front panel, or simply removed later. All incoming and outgoing power connections from the switches also share a ferrite bead which may reduce any high frequency electrical noise that the fan could put out. You could always install dedicated larger indicators or do more wraps of the bead from the fan power lines specifically or add dedicated inductors or film caps across those lines. Another idea would be to use a separate power source for the fan such as USB power from the capture PC if fan noise turns out to be significant. [I don't really know what I'm talking about here in terms of power rail noise reduction, but for now fan power rail noise is only a theoretical problem. I would still guess that the power rails have much lower electrical noise even with the fan on than a stock TBC-1000]

IMG_3851.jpg

You must be logged in to view this content; either login or register for the forum. The attached screen shots, before/after images, photos and graphics are created/posted for the benefit of site members. And you are invited to join our digital media community.

IMG_3852.jpg

You must be logged in to view this content; either login or register for the forum. The attached screen shots, before/after images, photos and graphics are created/posted for the benefit of site members. And you are invited to join our digital media community.

IMG_3853.jpg

[lordsmurf]

If you're going to mod it, then let's mod it. Don't screw around with fans. Headsink the chips and card. I forget the name of the "best" fan, but it's relative, as small fans are all pretty crappy.

Why the svid/comp choice? It can output both concurrent.

I'd like to see some sort of 3D printed grill. I hate gaping holes. Even taking some window screen material would be fine, fashion it to look nice. Maybe some window mesh + 3D printed frame.

Ferrite beads not overkill here.

From a quick glance, good solder work, maybe competent at worst.

Peel off those stickers, especially the "30224" that covers the VF-301x model (and I don't think it's even an OEM sticker).

I've seen a lot of TBC-1000 mods over the years. If ranked 1st (best) to 4th, then you'd get a strong 2nd for sure, based on appearance only. That's a decent mod job you've done, but you can do better, just keep going. Pay attention to performance, transparency, noise patterning.

[latreche34]

+1
I agree with heat sinks and an extraction fan mounted to to the back of the box blowing out, The suction port can be the rectangular hole on the front with some grille or wire mesh from a speaker enclosure. I personally despise old school toggle switches, I like modern square or rectangular rocker switches. The S-Video-Composite switch is not needed probably, the device senses the signal and switches accordingly internally, Better remove it (unless that capability is part of the dist. amp. which you've already removed), it adds an extra point to pickup noise into the video signal.

[aramkolt]

Appreciate the feedback for sure! My hesitation with heatsinks is that it is possible for them to detach/fall off if using thermal tape and permanently glueing them is iffy for me, though it's probably plenty reasonable to actually do. I do intend on removing the IC sticker as I agree that can be an insulator.

I believe the way the original configuration worked to "detect" which input was being used was a pin that gets physically bridge to ground when S-Video's round shield portion of the connector is inserted - so all the switch does here is bridge that contact to ground when S-Video input is selected here.

You'll notice that the original PCB has a much different looking outline (much larger) for the S-Video connector and that is to accommodate that "S-Video Detect" pin which is not present on a standard S-Video connector. I'll play around with it for sure to verify that that switch can't be deleted, but I'm pretty sure you wouldn't be able to toggle back and forth between without physically disconnecting cables without the switch.
With the switch, you could toggle back and forth to see if you're really getting much of a quality difference in the S-Video vs Composite coming off of the same video output device which would be useful to see if the comb filter inside of this does better than the one in your playback device.

You are correct though, both S-Video and Composite are simultaneously output without a switch and this switch doesn't affect that.

Main drawback to toggle switches is that they stick out and theoretically could get damaged, but I personally like click they make and find them to be quite reliable.

I do need to do some actual testing with an oscilloscope to see if there is any visible increase in electrical noise going into the TBC-100 card with the fan on vs off - having the switch for the fan for now will make that really easy to test.

Airflow is also very minimal here, so I don't think too much dust will get in for regular environments - but I agree, screens would be ideal on an "ultimate" mod, so I will probably add them as well. eventually.

I'll report back as far as what the fan did in terms of electrical noise once I do that oscilloscope testing and I'll report back on the voltage ripple of the Studer 900 power supply which should be pretty easy to measure, most likely in a week or two. I don't think there will be any visible "fan related image noise" mainly because the almost certainly noisier original power supply didn't show any noise with the fan on or off, but even knowing electrical noise is there would be enough to convince me to remove the fan or power it separately probably via USB. Hooking up a hardware vectorscope is typically my go-to for visualizing image noise as it shows up most easily there, even if you can't really "see" more noise.

Unfortunately, I don't have a un-modified TBC-1000 to compare "signal to noise ratio" of this modified version versus stock - Pretty sure the Tektronix VM-700T can report that statistic and I've got one that's mostly taking up space haha.

I'll post a capture of the output of a test pattern with the fan on vs off as well sometime in the next few weeks too.

[lordsmurf]

Why would thermal tape losing when the card sites flat? Buy better tape.

I don't care about toggles. I don't necessarily like rockers either. But I will say that toggles on the Elite BVP4+ tend to get damaged, so I do somewhat shy from toggles. I did some LiteOn DVD recorder modding in the 2000s, and added toggles for fans. I didn't run the fan at all times, just when using the LiteOn for marathon recording, as the DVD drives and chipsets would heat up too much.

If I left a big hole on my gear, and the my kitty got in the room, it would be a place for curious paws.

Sometimes noises is not cause by the power source. So it's possible for the fan to emit noise, even powered externally.

I may put a TBC-1000 in the marketplace soon, for a discount, because it really needs some bypass work, some cosmetic work, and I can no longer do it with my wrist injury. I'd rather sell it to somebody that can use it, rather than sit on it for 3-6 months while I heal. And those funds will directly go to the upgrade costs needed for this site. I'll have exact details on the listing, but it may serve you well for a comparison unit.