Thread
Terminating cheapo SCA -> 50Pin SCSI adapters
i have good news~!
Good news? the Chinese sellers will have the old adapters back soon?
Dear dreammacys,
hey i see you are selling these 68 to LVD…
i need SCA 68pin to SCSI 50pin…
you used to sell them just a few months ago…
what happened? i need like 50!!!
please let me know!
- morestuffuniwho
yeah man get those in… thanks… you were my guy… the only guy on the planet selling these for a reasonable price.
not only me but i know of a guy in the UK that buys these from you and ADDS termination with some caps and resistors
because they do not naturally come with termination. i don't thin 9 cents is going to effect peoples buying habits of them.
ThankYou!
- morestuffuniwho
hey i see you are selling these 68 to LVD…
i need SCA 68pin to SCSI 50pin…
you used to sell them just a few months ago…
what happened? i need like 50!!!
please let me know!
- morestuffuniwho
Dear dreammacys,
yeah man get those in… thanks… you were my guy… the only guy on the planet selling these for a reasonable price.
not only me but i know of a guy in the UK that buys these from you and ADDS termination with some caps and resistors
because they do not naturally come with termination. i don't thin 9 cents is going to effect peoples buying habits of them.
ThankYou!
- morestuffuniwho
Brilliant 
I'll order some more of them then
i'm gonna order some of these PCB's anyway, might as well now i have them drawn up
I'll order some more of them then
i'm gonna order some of these PCB's anyway, might as well now i have them drawn up
I've not used itead, but I've used Seeed and Smart Prototyping to make PCBs, and been mostly happy with both.
GKO is apparently the keep-out area, though I don't remember seeing it generated when I make a board with Eagle. See http://www.4pcb.com/pcb-file-submissions.html
Yeah seems the .GKO is supposed to have the keep out layer in it
However this http://digital-diy.com/forum/eagle/gerber-board-outline-layer-t2158.html says that what ITEAD actually want in the file is the dimensions layer
It seems it may not be needed, but i'll include it anyway
Also seems i need to move the outline of the board to the dimensions layer.
However this http://digital-diy.com/forum/eagle/gerber-board-outline-layer-t2158.html says that what ITEAD actually want in the file is the dimensions layer
It seems it may not be needed, but i'll include it anyway
Also seems i need to move the outline of the board to the dimensions layer.
Placed the order, so i guess i'll be seeing if the gerbers are formatted right soon (i can't think why there would be any problem with them)
Now... back to seeing if i can find cheap SCA connectors, don't really want to harvest them from chineese boards of i can avoid it
Now... back to seeing if i can find cheap SCA connectors, don't really want to harvest them from chineese boards of i can avoid it
yeah i dont blame you…
i am sure they get them from a distributor over there.
there is a phone number on this SCA to 50 pin adaptor i have here
tel 13316907839
- maybe you could call them and as to buy just some SCA connectors… if no…. then ask where they get them from.
never hurts to ask.
even jobs said him self said:
i am sure they get them from a distributor over there.
there is a phone number on this SCA to 50 pin adaptor i have here
tel 13316907839
- maybe you could call them and as to buy just some SCA connectors… if no…. then ask where they get them from.
never hurts to ask.
even jobs said him self said:
heh, maybe i should
i'm just gonna use the one i found in the uk for the first board, once i know the pcb's are correct i'll do some more digging
It's much easyier to search for them now i have found part numbers or a couple of parts
i'm just gonna use the one i found in the uk for the first board, once i know the pcb's are correct i'll do some more digging
It's much easyier to search for them now i have found part numbers or a couple of parts
Two thoughts, and I wish I had seen this before you ordered boards.
First, you might email your supplier of SCA to 50 pin adapters and ask him if he can get the components unassembled. You need everything except the circuit board anyway, and it should save him money with his supplier if they just ship the unassembled components.
I don't know if their flow will allow that to work, but it is another avenue, in case you can't get to a SCA connector manufacturer, or in case they want too much volume.
Second, would you mind explaining your termination scheme? It looks to me like you're tieing Term. Pwr. to the signal lines through 110 ohm resistors with a couple of diodes in the line to step down the voltage. I see what you're trying to do, but it is not going to be reliable. According to the literature, you can't count on Term. Pwr. being a reliable 5V. Plus, your voltage seen at the signal pin isn't constant because it depends on the current that flows through the single resistor -- no connection to ground, and that will change as the signal switches between 0 and 1. I'm willing to believe that I've overlooked some trick of electronics that makes this a reliable termination method, but it looks like a contraption for wildly varying termination current/impedance to me.
If one is going to tie the signal pins to Term. Pwr through 110 ohms of resistance, then the Term. Pwr. needs to be a voltage regulated source of 2.85V. This is active termination and it requires an actual voltage regulator in the circuit, such as an LT1086CT.
If you're not going to use a voltage regulator, then the old method of passive termination is to tie the signal pins to ground through 330 ohms and to Term. Pwr. through 220 ohms. This still has the problem that Term. Pwr. may not actually be 5V, but it has the virtue that the voltage at the signal pin is constant because of the voltage divider created between Term. Pwr and ground.
First, you might email your supplier of SCA to 50 pin adapters and ask him if he can get the components unassembled. You need everything except the circuit board anyway, and it should save him money with his supplier if they just ship the unassembled components.
I don't know if their flow will allow that to work, but it is another avenue, in case you can't get to a SCA connector manufacturer, or in case they want too much volume.
Second, would you mind explaining your termination scheme? It looks to me like you're tieing Term. Pwr. to the signal lines through 110 ohm resistors with a couple of diodes in the line to step down the voltage. I see what you're trying to do, but it is not going to be reliable. According to the literature, you can't count on Term. Pwr. being a reliable 5V. Plus, your voltage seen at the signal pin isn't constant because it depends on the current that flows through the single resistor -- no connection to ground, and that will change as the signal switches between 0 and 1. I'm willing to believe that I've overlooked some trick of electronics that makes this a reliable termination method, but it looks like a contraption for wildly varying termination current/impedance to me.
If one is going to tie the signal pins to Term. Pwr through 110 ohms of resistance, then the Term. Pwr. needs to be a voltage regulated source of 2.85V. This is active termination and it requires an actual voltage regulator in the circuit, such as an LT1086CT.
If you're not going to use a voltage regulator, then the old method of passive termination is to tie the signal pins to ground through 330 ohms and to Term. Pwr. through 220 ohms. This still has the problem that Term. Pwr. may not actually be 5V, but it has the virtue that the voltage at the signal pin is constant because of the voltage divider created between Term. Pwr and ground.
Interesting idea on asking for the boards unassembled, since that would give me all the parts. anyway, i'll work more on this once i have the pcb's, the ones i found in the uk are cheap enough for me to do initial testing.
How it is terminated is tieing 5v to the scsi lines through either 110ohm or 150ohm resistors (depending on if i'm using smd's or repacks, so in the case of these it will be 150ohm) plus 2 diodes (i'm using 1n4001's)
I did initially try using a 3.3v voltage regulator but just couldn't ever get it to work
i'll build a quick one using the active termination method at the weekend and see if i can get it working.
How it is terminated is tieing 5v to the scsi lines through either 110ohm or 150ohm resistors (depending on if i'm using smd's or repacks, so in the case of these it will be 150ohm) plus 2 diodes (i'm using 1n4001's)
I did initially try using a 3.3v voltage regulator but just couldn't ever get it to work
i'll build a quick one using the active termination method at the weekend and see if i can get it working.
Everybody keep on the lookout for the dumpster ballast level Sun SPARCstation 5 for some real goodies inside. There's a 2 SCA Drive backplane and cable in there with that's probably one of the very best Active Termination setups ever made! [
] ]'>
After I drill out a bunch of spot welds I'll post a thread about this backplane bonanza. I've (almost literally) been sitting on this Pile-O-SPARCstations I picked up at something like $5 ea. I was filching the purple blank and FDD Bezels out of one of the three 5s in the stack for use in a Sun Tape Drive Case and discovered this gem by accident this morning.
] ]'>After I drill out a bunch of spot welds I'll post a thread about this backplane bonanza. I've (almost literally) been sitting on this Pile-O-SPARCstations I picked up at something like $5 ea. I was filching the purple blank and FDD Bezels out of one of the three 5s in the stack for use in a Sun Tape Drive Case and discovered this gem by accident this morning.
That'd defiantly be interesting to see, do you think it is something that could be replicated in the limited space of the SCA though?
i'll be doing some playing with vreg's after collage tommrow, hopefully i still have some lm317's around, if i can get that working i will switch to using that
i'll be doing some playing with vreg's after collage tommrow, hopefully i still have some lm317's around, if i can get that working i will switch to using that
I was thinking more in terms of using the entire assembly in a larger Mac, but I'll get you some clear shots of the components/PCB for development purposes.
Ok, i really don't know why i couldn't get this working in the past
Used a bit of a crude setup to do some testing
[attachment=0]photo(20) copy.jpg[/attachment]
I couldn't get exactly 2.85v, but i tried a couple of differen't setups with the lm317, one gave 3.03v and one gave 2.5v. Both worked successfully
Then i tried a 3.3v regulator i had previously tried and that had failed, that also worked
So from now on i will use voltage regulators in all new builds
Used a bit of a crude setup to do some testing
[attachment=0]photo(20) copy.jpg[/attachment]
I couldn't get exactly 2.85v, but i tried a couple of differen't setups with the lm317, one gave 3.03v and one gave 2.5v. Both worked successfully
Then i tried a 3.3v regulator i had previously tried and that had failed, that also worked
So from now on i will use voltage regulators in all new builds
I like that setup, it's right in tune with my living room decor, ATM. Mac shui! ;D
Whats with the jellybean 78XX regulator? You might have a better streamline/crammage if you chose the 78LXX regulators, which come in a TO-92. Or, swap with a 1117 which comes in a SOT-89.
i have been using your adaptors in many different mac's and they have been working out fantastic for me… not one issue.
isn't this what were doing here, passive termination?
http://computer.howstuffworks.com/scsi4.htm
isn't this what were doing here, passive termination?
http://computer.howstuffworks.com/scsi4.htm
i love simplicity. if it works don't fix it…
The big reg is a LM317, just there for testing, will be using LM1117's for the actual adapter
The way they are terminated is actually closer to active termination, as trag mentioned earlier in the thread passive termination entails using two sets of resistors
And there isn't anything wrong with the way i have been doing it, just it relies on the 5v rail being the correct voltage, so if they where used in a computer with a wacked out 5v rail there is a chance you could get some damage (but if you have a wacked out 5v rail your probably gonna damage something anyway)
The regulators don't add any extra cost to the adapter, there a couple of pence each from china, + a tiny square of pcb (which i'll produce myself)
The way they are terminated is actually closer to active termination, as trag mentioned earlier in the thread passive termination entails using two sets of resistors
And there isn't anything wrong with the way i have been doing it, just it relies on the 5v rail being the correct voltage, so if they where used in a computer with a wacked out 5v rail there is a chance you could get some damage (but if you have a wacked out 5v rail your probably gonna damage something anyway)
The regulators don't add any extra cost to the adapter, there a couple of pence each from china, + a tiny square of pcb (which i'll produce myself)
yeah i am pretty sure BMOW has this very same issue with his Floppy Emu, The screen goes weird every time plug it into this 512k i have here, The thing works but might have some dirty power output or something.
maybe i should ask em to put a nice volt meter on the screen of his floppy emu??? that would be useful… pop on and you can see right whats/what. make it more useful and more of a diagnostic tool as well!
maybe just a one liner at the top? +12=XX.XX -12=XX.XX +5=XX.XX
. 5 -12V
. 6 +5V
. 7 +12V
. 8 +12V
or go the extra mile and program in limits… and the floppy emu says… HEY YOUR PSU IS NOT GOOD DOOD! FIX IT! hahah that would be Cool!!
maybe i should ask em to put a nice volt meter on the screen of his floppy emu??? that would be useful… pop on and you can see right whats/what. make it more useful and more of a diagnostic tool as well!
maybe just a one liner at the top? +12=XX.XX -12=XX.XX +5=XX.XX
. 5 -12V
. 6 +5V
. 7 +12V
. 8 +12V
or go the extra mile and program in limits… and the floppy emu says… HEY YOUR PSU IS NOT GOOD DOOD! FIX IT! hahah that would be Cool!!
There's a simple test is you think you have dirty power on the 512k, grab some big caps (1000uf or so) and stick them on the voltage rails on the floppyemu, should give it much cleaner power
Had the same issue with a project i built in my electronics class in school a year or so back, microcontroller went crazy every time i tried to power up the board, after a week of literally checking everything on the board i tried putting a 1000uf cap over the power input, and it worked perfectly from then on. Turns out the power brick i was using wasn't so well regulated
Had the same issue with a project i built in my electronics class in school a year or so back, microcontroller went crazy every time i tried to power up the board, after a week of literally checking everything on the board i tried putting a 1000uf cap over the power input, and it worked perfectly from then on. Turns out the power brick i was using wasn't so well regulated
Ok, this seems to work nicely, just 3 SMT components. This also removes the need to add the 10uf electrolytic cap
[attachment=0]photo(21).JPG[/attachment]
And it should cram into the custom PCB's fine aswell
[attachment=0]photo(21).JPG[/attachment]
And it should cram into the custom PCB's fine aswell
Actually, I think there was.
With a voltage regulator in there, as you've switched to, when the signal pins are high (2.85 - 5V) there will be some current flow from the signal pin through the 110 ohm resistor to the regulated voltage area, damping ringing on the high end.
Under your previous configuration, when the signal was high, there could be no current flow to damp the ringing because the diodes acting as the voltage regulator would be reverse biased. Using diodes as a voltage regulator only works if the current flow is always in the same direction and if the difference in voltage between the ends is large enough to forward bias the diodes. Neither of these conditions are always met in this application.
I have not given full consideration to how the capacitor may have contributed to the termination, though, hence my earlier request for an explanation of how it was meant to work. It's possible I've overlooked an aspect which would make it work. But based on what I know so far, I really don't think it should work properly -- using a real voltage regulator is a much more robust solution.
Of course, in many instances, even no termination will still allow a SCSI bus to operate, so the fact that some systems with the old style terminators work is not proof that they're doing their job properly.
Anyway, all that said, I think it's great that you're moving to using a real voltage regulator circuit. Nice work.
Ahh i see. Well, it worked where no termination didn't, but isn't the correct way to do it
So, is it ok to use a 3.3v reg like the one i used in that picture? Been looking around and they are cheaper and easier to get than the 2.85v variant
So, is it ok to use a 3.3v reg like the one i used in that picture? Been looking around and they are cheaper and easier to get than the 2.85v variant
I do not know enough to answer this question. The specification says to use 2.85V. I imagine that you could get some ringing on the high signals because 3.3V is a little too high and your highs will be a little underdamped, and your lows will be a little overdamped. With slow original SCSI and for these applications, will it ever make a difference, or will it be "good enough"? I just don't know. There was probably a time when I could have calculated/analyzed it, but those days were about a decade ago....
If you can get 2.85V or an adjustable Vreg without blowing the budget, I'd go that route. I think the LM317 is adjustable, but then you need more components. You might hunt around at National Semiconductors, TI, and Linear Tech (I think someone bought NS a year or two ago) in their power and control or voltage regulation sections and identify a few simple chips that can do the job, then hunt them on Ebay to see the pricing. Also, take a look at questcomp.com. Sometimes they have stuff in stock at lower prices than Ebay. The LT1086 is a good choice, if available in a smaller package than I've seen it in, but I don't know about pricing. LT might have something similar with a lower current rating. Oh, it's worth checking Digikey and Mouser as well. Oh, you're in the UK right? So Farnell?
i'll just find some 2.85's then, better to do it correctley now than have to change it again further down the line
This is a good price for them : http://www.aliexpress.com/item/Ams1117-2-85-2-85v-1a-power-supply-voltage-stabilizing/771958640.html
I can't seem to find any uk suppliers that do them for a decent price, so it looks like getting them from china is my only choice
This is a good price for them : http://www.aliexpress.com/item/Ams1117-2-85-2-85v-1a-power-supply-voltage-stabilizing/771958640.html
I can't seem to find any uk suppliers that do them for a decent price, so it looks like getting them from china is my only choice
thats a good deal man
careful they might be fakes. In UK you wanna use farnell.
Yeah that was my worry, However i think i'm going to take the risk. As the cheapest i can find them is from in the uk is from RS, ~30 pence each (+ shipping +a card handling charge as my order would be under £20)
Edit: the best ones from farnell is : http://uk.farnell.com/rohm/bu2jtd3wg-tr/ldo-fixed-2-85v-0-2a-ssop-5/dp/2342945
However the £20 pretax minimum order limit is a real problem for me
Edit: the best ones from farnell is : http://uk.farnell.com/rohm/bu2jtd3wg-tr/ldo-fixed-2-85v-0-2a-ssop-5/dp/2342945
However the £20 pretax minimum order limit is a real problem for me
that website seems fairly massive.
I doubt a vendor would risk getting black listed from such a large sales medium...
but then again ... Ya never know... I would say its worth the 8-10 bucks.
I was going to buy 2000 Murata 16v 47uf ceramic caps from there last fall... for like 206.00 free shipping.
sick deal!
I have not found a deal like that this new year.
mouser also sells that brand.
I doubt a vendor would risk getting black listed from such a large sales medium...
but then again ... Ya never know... I would say its worth the 8-10 bucks.
I was going to buy 2000 Murata 16v 47uf ceramic caps from there last fall... for like 206.00 free shipping.
sick deal!
I have not found a deal like that this new year.
mouser also sells that brand.