Seagate ST-4038

[dkedrowitsch]

I know this is a long shot, but I have an ST-4038 here that I adopted some time ago. It was apparently handled rough and had a few components mangled and one busted off. Mostly just caps and one choke which were easily replaced, but the tricky one is actually completely missing. It's an inductor which I assume has something to do with the voice coil control. It's on the top board on the front right corner (circled in attached picture).

Would anyone happen to have one of these drives that's damaged who could send me the top board? Or maybe even just the part number of that inductor? None of the pictures on the internet are detailed enough to read the part number.

Or, if someone was willing (and has the tools) to unsolder the two legs and use an L/C meter I could use the inductance reading to wind a replacement coil of proper value myself (we have a coil winding machine at work).

Anyway, I know my chance of success is slim. But I figured I'd put this out there just in case.

By the way, the drive currently spins up and unlocks, but that's it. No head action at all. Before I repaired all the other damage, it would spin up and immediately after unlocking slam the head assembly against a stop.

 

[MikeS]

Seagate ST-4038

Well, I've got the MFM version here, the 4026; the PCB layout is somewhat different, but it does have a similar inductor in about the same location.

I'm not quite ready to scrap it yet (still working last time it was tested) and I'm not too crazy about removing & testing the coil, but FWIW the markings are:

H5A
5 X
A250

mike

 

[dkedrowitsch]

This is an MFM drive as well, only it's the 30meg (formatted) version of yours.

Thanks for the info, I'll see what I can dig up!

 

[Druid6900]

Ok, I have the bottom board of a 4038 and I'm pretty sure I have the top board around too (just requires some digging).

There is a sticker on it, in my handwriting that says "Bottom bad, top good", so, if I can find the top, it's yours for shipping.

I can't PROMISE I have it, but, I'll look around most thoroughly.

 

[dkedrowitsch]

SWEET!

PM me if you find it!

Ok, I have the bottom board of a 4038 and I'm pretty sure I have the top board around too (just requires some digging).

There is a sticker on it, in my handwriting that says "Bottom bad, top good", so, if I can find the top, it's yours for shipping.

I can't PROMISE I have it, but, I'll look around most thoroughly.

 

[geneb]

Many years ago (like 20) I actually combined two of those drives to get one working one. One drive supplied the electronics and the other drive donated a platter.

I was amazed it worked as long as it did. I opened the drive up and swapped out the top platter by hand. It formatted properly and didn't start pitching errors for about 4 months. After a while even Spinrite could save it.

Thanks for triggering the memory.

g.

 

[modem7]

Or maybe even just the part number of that inductor?

On my ST-4038, the coil is marked:

H28B
624

I don't have any way to measure the coil's inductance.

Looking at the ST-4026 though, I have three of those and on each one, the markings differ slightly, and they are completely different to what MikeS reported on the coil of his ST-4026.
And so the marking on the coil of my ST-4038 could be misleading (eg. does not contain a part number).

Between the ST-4038 and the ST-4026, that area of the top board is very similar:
1. 1 x subject coil
2. 1 x IC - uA78S40PC (switching regulator)
3. 1 x Choke
4. 1 x small electrolytic cap
5. 1 x 2N5023 transistor
6. Medium electrolytic caps (220uF on the ST-4038, two smaller caps on the ST-4036)

It looks like the ST4026 and ST-4038 might share that same small part of the circuitry, eg. the subject coil could be identical between them.

There is a datasheet for the uA78S40PC at http://www.datasheetarchive.com/pdf/3511263.pdf
In that datasheet, Figure 7 shows an example usage which fits with the above circuitry nicely, particularly when I can see that on the ST-4038 that one of the coil's leads is connected to the 2N5023 and to the cathode of a diode. And I can see (measure actually) that one leg of the 2N5023 goes to pin 14 of the uA78S40PC.

Therefore I have no doubt that Figure 7, and the table below it, is the key to determining a value for the coil (unless of course you can find someone to measure the actual value used by Seagate).

The table below Figure 7 specifies the formula to calculate the component values. Note that the description of the coil is "L(min)", i.e. a minimum value.
Figure 7 shows the coil as 50uH. Maybe you play safe and throw in something substantially larger - say a 200uH.

 

[dkedrowitsch]

Thanks for the info! I didn't even bother to look up the datasheet for the SMPS regulator, which was a great idea.

I did a little poking around last night and the H5A markings on MikeS's drive apparently indicates the properties of the ferrite core the coil is wrapped around, but doesn't indicate the inductance of the coil (at least that I can find). I think it's a TDK part #, and according to some of their PDFs they're intended for telecom and SMPS applications. Which means there may be more involved then just the inductance value alone. How sensitive the circuit is to variances in whatever other properties there are to these coils, I don't know. Maybe the switching regulator has a wide tolerance.

The fact that several of your drive's coils have different markings is actually encouraging...hopefully that indicates there is indeed a tolerance.

Since there is a switching regulator and high power transistor involved, it looks like this is probably a DC/DC converter which I assume is for a high current or voltage supply to the voice coil controller.

I have access to dozens of various SMD coils used in my company's products, specifically the SMPS section of the PCBs. I'm looking at one now which is a 2 winding Coiltronics with a parallel inductance of 200uH (138uH full load) and max current of 480mA.

I'll examine the datasheet you linked and see if it looks like I can easily substitute this part.

Or maybe I'll get lucky and Druid6900 will find his spare top board.

On my ST-4038, the coil is marked:

H28B
624

I don't have any way to measure the coil's inductance.

Looking at the ST-4026 though, I have three of those and on each one, the markings differ slightly, and they are completely different to what MikeS reported on the coil of his ST-4026.
And so the marking on the coil of my ST-4038 could be misleading (eg. does not contain a part number).

Between the ST-4038 and the ST-4026, that area of the top board is very similar:
1. 1 x subject coil
2. 1 x IC - uA78S40PC (switching regulator)
3. 1 x Choke
4. 1 x small electrolytic cap
5. 1 x 2N5023 transistor
6. Medium electrolytic caps (220uF on the ST-4038, two smaller caps on the ST-4036)

It looks like the ST4026 and ST-4038 might share that same small part of the circuitry, eg. the subject coil could be identical between them.

There is a datasheet for the uA78S40PC at http://www.datasheetarchive.com/pdf/3511263.pdf
In that datasheet, Figure 7 shows an example usage which fits with the above circuitry nicely, particularly when I can see that on the ST-4038 that one of the coil's leads is connected to the 2N5023 and to the cathode of a diode. And I can see (measure actually) that one leg of the 2N5023 goes to pin 14 of the uA78S40PC.

Therefore I have no doubt that Figure 7, and the table below it, is the key to determining a value for the coil (unless of course you can find someone to measure the actual value used by Seagate).

The table below Figure 7 specifies the formula to calculate the component values. Note that the description of the coil is "L(min)", i.e. a minimum value.
Figure 7 shows the coil as 50uH. Maybe you play safe and throw in something substantially larger - say a 200uH.

 

[MikeS]

This is an MFM drive as well, only it's the 30meg (formatted) version of yours.

Thanks for the info, I'll see what I can dig up!

Ah, learn something every day; I just assumed that the '38 was the RLL version of the '26, but apparently not.

I've got a dead 4096 or two but they don't use that inductor any longer; hope Richard can come through for ya.

Good luck!

mike

 

[Sharkonwheels]

I believe Seagate appended "R" to the model number on RLL versions, ie,
ST225R,ST238R, ST277R, ST4144R, etc... "N" appended was SCSI (maybe except the NV, which were the Pre-CompSurfed Novell drives), A=ATA, etc..

I have an ST-238R on a CoCo3, with a Burke & Burke XT/RTC interface, and (2) spares... got a few ST-251-1's (28ms seek) on some Kaypro 10's, and 1-2 extras (well, until the get installed

Now, ST-4096 I remember well...I remember having 2 of those beasts, in a 386SX-25, running on a PerStor controller, which turned each drive from 80MB to about 130-140MB. This was in about 1990-1991 or so. Over 200MB of storage was like unreal back then!!!!


T

 

[dkedrowitsch]

Well, I've thrown some various inductors at it ranging from 100-400 uH and while watching with an oscilloscope I do get a reaction, a nice clean 12v PWM signal, but the drive still isn't happy. After it spins up and unlocks, I simply get a nice hard slam against the stops.

Would someone mind taking a DC voltage reading between the drive ground and left leg of the coil and tell me what voltage they see? The same signal would be on the case of the large metal bodied transistor just north of the coil on the other side of regulator IC.

 

[modem7]

Would someone mind taking a DC voltage reading between the drive ground and left leg of the coil and tell me what voltage they see? The same signal would be on the case of the large metal bodied transistor just north of the coil on the other side of regulator IC.

+ 0.11 Vdc measured.
It's clear that this part of the circuitry is a +12V/-12V converter. The input, pin 13 of the IC, measures +12.1 Vdc. The output, anode of the diode, measures -11.78 Vdc.

 

[Druid6900]

Still looking for that top board.......

 

[MikeS]

Still looking for that top board.......

... and I'm sorry I dug out that 4026 to look at; turns out that it too has some components ripped off the top board, although it looks like they're all bypass caps tacked on after the fact. Just what I needed, another "project"...

And of course I knew about the "R" suffix for RLL; the aging gray cells just saw the "38" and leapt to the conclusion that it was like the 225/238... what can I say...

Get a little snow out your way, Druid?

mike

 

[dkedrowitsch]

Thanks, because it looks like I'm going to need it if you do find it.

During my experimentation with different inductors, something else on that board went on the blink and I'm not getting any oscillation from the regulator IC anymore.

I may be able to substitute everything in that circuit with a canned +/- 12V converter, but at this point I'd rather just stick another board on and be done with it.

In the meantime, I managed to dig up a nice working ST02 SCSI/Floppy controller in my box of parts and was running an old Maxtor LX 300something meg drive in the "beater" 5150 I got in my last thread when the power supply took a dump in the middle using the computer. It was the 65W model, so it's probably my fault since they're not supposed to have hard drives. Time to hack up another AT power supply....

Still looking for that top board.......

 

[Druid6900]

Get a little snow out your way, Druid?
mike

Nothing to write home about. 25, maybe 30 cm, just ran the old snow blower over it a couple of times

You get dusted?

Ok, the board has to be in the basement somewhere and I'm going to risk going down there to track it down.

If I don't make it, Mike can have the keyboard cable for the SD-550 motherboard so he can give up looking for it LMAO

Update: Found a top board, but it's for a ST-4051. It looks VERY similar, but, I don't know if I'd go using it for a direct replacement. I'll keep looking for the 4038, but, if I don't find it, this board may be good for pulling parts off of.

It's only a servo board, so, I don't think they would be too dissimilar.

 

[MikeS]

Nothing to write home about. 25, maybe 30 cm, just ran the old snow blower over it a couple of times

You get dusted?

Ok, the board has to be in the basement somewhere and I'm going to risk going down there to track it down.

If I don't make it, Mike can have the keyboard cable for the SD-550 motherboard so he can give up looking for it LMAO

Update: Found a top board, but it's for a ST-4051. It looks VERY similar, but, I don't know if I'd go using it for a direct replacement. I'll keep looking for the 4038, but, if I don't find it, this board may be good for pulling parts off of.

It's only a servo board, so, I don't think they would be too dissimilar.

Snowblower??? Feh! Real Men use Shovels!
But I'll make an exception in your case...

No, I told you, I finally found that cable and a dozen or so just like it, so if you ever need another one...
- glad you survived though.

Probably not relevant here, but I've replaced a board on a Maxtor with one from a different model and to my delight and surprise it worked just fine even though the geometry was different.

mike

 

[dkedrowitsch]

GREAT! If you don't find the 4038, let me know what you want for shipping for the board you have. I'll PM you my address so you can get a price.

THANKS!

Update: Found a top board, but it's for a ST-4051. It looks VERY similar, but, I don't know if I'd go using it for a direct replacement. I'll keep looking for the 4038, but, if I don't find it, this board may be good for pulling parts off of.

It's only a servo board, so, I don't think they would be too dissimilar.

 

[MikeS]

GREAT! If you don't find the 4038, let me know what you want for shipping for the board you have. I'll PM you my address so you can get a price.

THANKS!

While we're at it, Druid, you're the man to ask:

DK was asking me about a monochrome monitor; with your experience in these matters, any ballpark idea how much it would cost to ship one to PA (assuming he's still interested)?

mike

 

[Druid6900]

While we're at it, Druid, you're the man to ask:

DK was asking me about a monochrome monitor; with your experience in these matters, any ballpark idea how much it would cost to ship one to PA (assuming he's still interested)?

mike

Last monitor I shipped on its own was to New Jersey and it was the colour monitor from a PC jr. I think it cost me about 40 bucks CAD to ship (packed to arrive alive, which it did), but that was about a year ago.

A mono monitor would, most likely, weight less than any IBM monitor (especially a colour one), so, taking into account the increase in CPC's prices, I'd guess in the 30 to 40 dollar range using Expediated Parcel (includes $100 insurance and is trackable)