Image Map Image Map
Page 1 of 2 12 LastLast
Results 1 to 10 of 19

Thread: Replacement transformer recommendation fro Northstar Horizon ?

  1. #1
    Join Date
    Nov 2017
    Location
    UK, South East
    Posts
    115

    Default Replacement transformer recommendation fro Northstar Horizon ?

    Can anyone recommend a replacement transformer for the Northstar Horizon please ?

    Mine seems to be producing around 20V and 10V rather than 16V and 8V, regulators on both memory boards are uncomfortable to touch (hot!) and it appears that over time the heat has discoloured the PCB around the heatsink mount (perhaps this is normal?) so I'm looking to try and get things back closer to spec.

    I did look at the article below where another owner has done similar but the spec of that transformer does not look correct.

    http://www.connect.gi/northstar.htm

  2. #2
    Join Date
    Jan 2007
    Location
    Pacific Northwest, USA
    Posts
    28,268
    Blog Entries
    20

    Default

    Transformers for S100 systems of that genre were usually custom jobs (not expensive back then if you wanted to order 100 or 1000 of them).

    Instead of replacing the transformer, consider adding a small one in the primary circuit as a "buck" source. To elaborate: you want to reduce the output voltage by 20%, so reduce the input voltage by 20%, which would be 0.2 x 120V = 24V. So power a 24VAC output transformer from the line and hook the secondary in series with the main power transformer primary, so it "bucks" the AC input. Note that the output current rating only needs to match the primary current rating of your main transformer.

    Just an idea...

  3. #3
    Join Date
    Dec 2005
    Location
    Toronto ON Canada
    Posts
    6,962

    Default

    What do you think is wrong with the specs?

    I assume the transformer doesn't have a multi-tap primary?

    I've got a few transformers from scrapped systems but the shipping would be a killer.

    Chuck's idea sounds like the way to go, although it looks like you'd need a somewhat harder to find 48V transformer (assuming we're talking 240V); if nothing else connects to the primary (8" rives etc.) you could probably even put it in an external box.

    m

  4. #4
    Join Date
    Nov 2017
    Location
    UK, South East
    Posts
    115

    Default

    Its a good idea Chuck - thanks. Not sure I have the physical space to fit another transformer in the case.
    I presume you meant find a transformer with 120V-(120*0.2) = 96V output in your example or 192V for me as I'm in the UK using a 240V supply ?

    EDIT: Found this - seems to explain the 'buck' approach more clearly to me:

    http://sound.whsites.net/articles/buck-xfmr.htm

    This was quite an interesting note I found :

    http://www.retrotechnology.com/herbs_stuff/s_power.html

    Seems to imply my situation is quite common with (lightly loaded) Horizons and suggests a few other ways to reduce the voltages including variac or 'buck' mechanism
    Last edited by zippysticks; May 4th, 2018 at 09:44 AM.

  5. #5
    Join Date
    Nov 2017
    Location
    UK, South East
    Posts
    115

    Default

    Hi Mike - really just that they are 20% over spec - although my system is lightly loaded with just CPU, memory, floppy controller and twin floppy drives.
    Despite already lasting 40 years, I don't like the regulators running that hot on the memory cards.

  6. #6
    Join Date
    Jan 2007
    Location
    Pacific Northwest, USA
    Posts
    28,268
    Blog Entries
    20

    Default

    There are a couple more ways to skin this cat. You could simply insert a fixed voltage drop with a couple of power rectifiers in series (0.7V per diode) or use an LDO regulator (e.g. LM1084-ADJ, which is rated at 5A). Power transformers are usually designed to deliver the nameplate voltage at a specified current. For example, I've got a couple of "control" transformers here that are rated for 12.6V RMS at 4A. No-load on these is closer to 15VRMS. No matter what you use, absent some sort of switching regulator, any llinear approach that drops the DC voltage directly will dissipate heat. So, to drop 2V at 5A, you're going to need to get rid of 10W of heat.

    The other issue is that with a simple capacitor-input filter, the no-load output approaches the peak-to-peak AC voltage output of the transformer, which is about 1.414 x the nameplate RMS voltage.

    Another approach would be to replace the 78xx linear regulators with 3-terminal drop-in switching regulators. There, the input voltage doesn't make as much of a difference, since the regulator only draws what's needed to maintain the output. But that would entail replacing the on-board regulator(s).
    Last edited by Chuck(G); May 4th, 2018 at 10:45 AM.

  7. #7
    Join Date
    Dec 2005
    Location
    Toronto ON Canada
    Posts
    6,962

    Default

    Quote Originally Posted by zippysticks View Post
    Hi Mike - really just that they are 20% over spec - although my system is lightly loaded with just CPU, memory, floppy controller and twin floppy drives.
    ...
    Ah, OK; I thought you meant the replacement: 8.5 and +- 17.5 sounds reasonable.

    Wonder where he got it; does he mention it anywhere?

    m

  8. #8

    Default

    Thanks for the compliments about my notes on that S-100 Web page. I ran into similar problems on an H-89 I was restoring: DC voltages too high due to AC primary voltage too high; and marginal-voltage tantalum caps. I've added a link from that S-100 Web page, to my H-89 notes at:

    http://www.retrotechnology.com/restore/h89_4th.html

    The general problem with excessive unregulated DC voltages on vintage 70's computers, seems to be that modern AC line voltage is higher, near or above 120V AC, than AC line voltage decades ago. Solutions have been discussed in this vcfed thread; my two Web page go through some of the details. I suppose I should explain a bucking transformer configuration, the details aren't obvious, Lee Hart explained them to me.

    Herb Johnson

  9. #9

    Default

    Quote Originally Posted by zippysticks View Post
    Its a good idea Chuck - thanks. Not sure I have the physical space to fit another transformer in the case.

    EDIT: Found this - seems to explain the 'buck' approach more clearly to me:

    http://sound.whsites.net/articles/buck-xfmr.htm
    Wow thanks for the link, figure 4 looks to be the way to go. A 230V to 18V (9V + 9V) or 24V (12V + 12V) transformer should be easily available. I would suggest mounting this in a separate box with a fuse on the input, and feed the output to a chassis socket. The only caveat I would have is a lot of transformers are not marked with a dot, so you would need to power it on with nothing connected and check the output voltage is lower than 230V. If wired incorrectly the output would be higher.

    Regards,
    John

  10. #10
    Join Date
    Apr 2015
    Location
    Austin, Texas
    Posts
    1,234

    Default

    Quote Originally Posted by hjohnson View Post
    The general problem with excessive unregulated DC voltages on vintage 70's computers, seems to be that modern AC line voltage is higher, near or above 120V AC, than AC line voltage decades ago.
    While the mains voltage being higher is a factor, it's not the primary cause of transformer voltage drift.

    The windings start shorting over time as the varnish on the copper migrates or deteriorates, which can be accelerated by excessive heat. The single core wound types where the primary/secondary are wound on top of each other tend to have worse problems with it, while the dual bobbin type have less since they dissipate heat better and aren't generally wound as tightly. The shorts can happen on both the primary side and the secondary side, I've pulled dozens of transformers apart which had primary, secondary or both which had shorts. When you start unwinding the copper wire, the varnish will be irregular in glops or waves with often many exposed section of copper.

    This shorting process can happen very quickly, I've had transformers not a year old measure 20% out of spec and it only gets worse with time. The worst transformer drift case I had was a 9v for a game console that drifted upwards in excess of 24v, itself being about 27 years old. This transformer hadn't been in use most of that time either, which shows it doesn't need to be used to drift.

    It's always a good idea to check linear transformers before you use them so they don't nuke something with too much voltage.

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •