Jeffrey LewisHasaka Spinning Factory Revisited

Earlier this week, I posted about a location in Syria that some folks in the IAEA believe was originally intended to house centrifuges as part of a nuclear weapons program.

This is just a housekeeping post, bringing that conversation up to date based on comments and few off-line conversations. The short version is that it is almost certainly a textile factory today.  The allegation, apparently, is that the facility was intended to house centrifuges.  Who knows?

First, we now have two bits of documentation in Arabic located by one of our students, Anna Eorkhina. (One of the best parts about MIIS are all the talented young people with amazing language skills.)

She found, posted on-line, a pair of stories in Arabic about the Hasaka Spinning Factory.

A daily publication on local politics in Hasaka carried a March 5, 2009 story titled Benefits and New Features of the Hasaka Spinning Company that notes a Presidential Decree (No, 89, 12 February 2009) established the financial and administrative independence of firm, which had previously been part of a state entity.  The General Manager (pictured above) explained that the firm has existed since 1976, but could not be considered a company until now.

Another site, which looks less professional, carried a November 15, 2009 story with some additional details about the timeline of the firm: Another reference to the founding in 1976, machines imported from East Germany in 1983, a major modernization from 2001-2008.

In the comments, a reader named “Mark” confirmed that “some people” believe the plant was originally intended to house centrifuges. (I’ll vouch for him.) I have also heard some hints about what might be the cause of the suspicion, but I’ll hold off on that for now.  The important point, of course, is that no one is disputing that the facility is, now, really an honest-to-goodness textile factory.

Andy, one of our most frequent and insightful commentators, noted a few things including the lack of high-voltage lines that would be necessary for either a textile factory or a centrifuge plant.  I wonder if they might not be buried, which would be an excellent sign of concealment and deception.  I have to say, can you imagine being some little state-run firm, getting this amazing new facility and then walking around like “Wow, that’s a really big scale in the floor …”

Finally, one reader — in mocking the notion that the facility might be anything other than a textile factory — noted the date of the DNS registration for the site was 2007-12-7.  I found a January 2007 date using another query, but the basic question is whether the website went up after Israel demonstrated that the covert Syrian nuclear program had been compromised in September 2007.

Since it appears the allegation involved the original intent of the facility, as far as I can tell we are working on two questions: Does the construction timeline for the facility square with the timeline given for the Hasaka Spinning Factory and does the facility have signatures that would suggest it was built for some other purpose?


  1. Andy (History)


    I guess I had assumed that this place was one location on the list of possible places where Syria took some of the material and equipment from al Kibar after it was bombed. There were a few other facilities similarly suspected of being storage sites. But that’s not the reason for suspicion here – it is a suspected (prior) enrichment site instead?

    I haven’t heard anything about centrifuges related to Syria – is there anything else besides what’s been alleged here?

  2. Hairs (History)

    Minor quibble alert:

    “Andy, one of our most frequent and insightful commentators, noted a few things including the lack of high-voltage lines that would be necessary for either a textile factory or a centrifuge plant. I wonder if they might not be buried, which would be an excellent sign of concealment and deception.”

    Underground transmission cables have much higheer harmonics than overhead cables, and harmonics are something you’d be trying very hard to avoid in a centrifuge plant – which is why low-harmonic frequency converters above about 600 Hz are restricted for export.

    Given the importance of a low-harmonic supply to the centrifuges, in my opinion any putative enricher is more likely to bring in his power via overhead cables (and disguise the purpose by building, say, a small smelting plant or other excusable consumer e.g. textile factory!) than deliberately “dirty” his power supply and then have to spend extra resources and effort cleaning it up again on yon side of the frequency converter.

    In this respect, the absence of visible power cables at a suspect facility is marginally more a sign of innocence than guilt.

    • Jeffrey (History)

      There are, however, high voltage power lines nearby. How close would a building containing the conversion equipment need to be?

    • Hairs (History)


      I don’t think there’s any absolute distance that we could say is / isn’t too far for the supply of centrifuges from underground transmission lines.

      The key issue is that transmission cables below ground experience greater capacitance and inductance than overhead cables. They also experience greater changes in capacitance and inductance as a result of differences in the geology / geochemistry along their length, and also great changes associated with changing soil moisture and water table height. The net result of all of this is that (in comparison to overhead transmission) the further the cables run below ground the more likely they are to develop unwanted harmonics.

      Of course, harmonics can be suppressed or removed by suitable filtering, but the maxim that prevention is better than cure applies as much in engineering as anywhere else. My guess – and it is only a guess – is that enrichers would prefer to get the “cleanest” possible power supply, and disguise what’s going on in other ways, than allow the supply to become sullied and thus reduce even by a small amount both the integrity and efficiency of the operation.

    • Ano N. Ymous (History)

      Doesn’t desire to have “cleanest” possible power and minimize harmonics imply DC power? Especially since any semi-modern frequency converters (GTO and on) use a DC mid-step anyway. If one used DC-fed freq. converters and a single big rectifier, one should be able to afford much better filters at that one rectifier unit than for hundreds of individual, smaller yet not consumer electronics class rectifiers.

      RF reception would still happen of course, unless properly shielded against, but a wire does that regardless of whether there’s DC or AC flowing through.

    • Hairs (History)

      Ano N. Ymous:

      I don’t *know* how frequency conversion is done for centrifuge plants, although I suspect you’re correct that it’s via an interim DC section i.e. via the gate turn-off thyristors (GTOs) that you mention or possibly via the integrated gate commutated thyristors (ICGTs) that are becoming more popular for high power variable frequency applications in the power industry.

      But the adage that the quality of an output rarely exceeds the quality of the input still applies: once GTOs start switching above about 800 Hz then they suffer increasingly from residual currents because the residual charge can’t be instantaneously removed, and while ICGTs will switch faster than GTOs their losses are also higher, which I assume would be undesirable for centrifuge applications.

      Whatever the scheme used, the quality and stability of the power supply to centrifuges must be a key factor in the stability and reliability of their operation (why else would Stuxnet target the speed controllers?), and since harmonics represent “jitters” in a centrifuge’s speed it seems logical to me that the operators would want harmonic free supply (again, it is only low-harmonic frequency converters that are export regulated – trashy high-speed converters can apparently be freely traded). Since no speed conversion technology is completely free of foibles (e.g. GTOs’ residual currents) I’m inclined to think that if a centrifuge operator can avoid harmonics at the input to his power supply then he’d be inclined to do so, even at the cost of having overhead transmission cables coming right up to his door – something which he’d have to explain away in another fashion.

  3. blowback (History)

    “but the basic question is whether the website went up after Israel demonstrated that the covert Syrian nuclear program had been compromised in September 2007”

    What evidence is there that Syria had a covert nuclear program in 2007? If you use the “criminal” test (beyond reasonable doubt), there is none. All we really know for certain is that there was a structure near the Euphrates which was demolished/destroyed and replaced by another structure and that traces of anthropogenic uranium were subsequently found at that unsecured location. If we are to believe Bob Woodward’s account, the only evidence that the US government had came from the Israelis and they are not exactly known for their generosity with the truth. As for the presentation given by the US government after the alleged event, I need hardly remind anyone of a certain “important” presentation by the US Secretary of State to the UN in 2003 which, let’s be honest, was complete garbage.

    • Jeffrey (History)

      Yes, Powell’s presentation was complete garbage, but I don’t see any reason to dispute the evidence that the building at Al Kibar housed a reactor.

  4. rwendland (History)

    While the domain registration of was in 2007, the website content appears older, from 2006.

    The copyright notice is “(c) 2006 Hassakeh Spinning Project”, and if you download photos from that website and look at the EXIF data in the JPG (normally put in by the camera) you get 2006 (just) as well.

    While of course this could be faked, it does suggest the website is from well before Israel bombed the claimed reactor.

    For example the JPG EXIF data for this photo of inside the textile factory:


    Camera make : Mercury Peripherals Inc.
    Camera model : DigitalCam Pro
    Date/Time : 2006:12:29 14:31:39
    Resolution : 1024 x 768
    Flash used : No
    Focal length : 7.5mm
    Exposure time: 0.250 s (1/4)
    Aperture : f/2.8
    ISO equiv. : 68
    Metering Mode: center weight
    Exposure : program (auto)
    Jpeg process : Baseline

    Also, a Syrian govt website claims the “Hassakeh spinning project established in 1983”, well predating the copyright date used on the website.

  5. kme (History)

    Well, if the building now housing the Hasaka Spinning Factory was once intended to house centrifuges, then we can say one thing for sure: that the Syrians certainly have a droll sense of humour.

  6. b (History)

    As for the establishment of that company, keep in mind that many Syrian factories have been, until recently, state owned and run within a socialist system.

    (Slow) preparations for privatization, like establishing factories as real companies, only started under Bashir Assad after 2000.

    Same thought on the website of that company. Under a socialist system it did not need any. The central plans took care of production quotas and customers. Only the perspective of a system change and eventual privatization would make a website necessary.

  7. bradley laing (History)

    –Is every industrial facility in Syria a potential target for Israeli jets? During WW II, Warner Brothers Studio put a sign on the roof of their buildings saying “Lockheed this way,” so that any Japanese bombing planes would not mistake WB for an airplane factory.

    —Also, note that the paranoid Albanian government filled the countryside with obsolete military bunkers, ones thqat faced both towards potential invaders, and towards the population of the cities themselves. Could the Syrian architects simply be living in paranoia-land?

    • Smith (History)

      “During WW II, Warner Brothers Studio put a sign on the roof of their buildings saying “Lockheed this way,” so that any Japanese bombing planes would not mistake WB for an airplane factory.”

      From what I gather, this is not true. Warner camouflaged their studio a la Lockheed, as did at least one other regional airport, but all signs (snicker) point towards there being no sign directing them to Lockheed.

    • Ano N. Ymous (History)

      I think any such sign would have been altered to read just “Lockheed” within seconds from being spotted from any of the myriad US military planes flying at the west coast at the time. Point being, which facility do you think contributed more to the US war effort? Certainly luring the bombers to hit the less important facility would have been seen as a smart tactic.

    • Bradley Laing (History)

      —-At least one website claims the exact opposit of “Lockheed this way” being placed on the Warner Brothers Studio happened.

      —My example is suspect. But what about the threat of Israeli airstrikes? Is there an imagined, or even real, fear that the Israelis will just bomb Syrian randomly, in Syria?

  8. Anon (History)

    The Syrians ought to do their cotton fabrication in underground bunkers!

    UNSCR 487 is one of the UNSC resolutions that is selectively deleted from collective memory.

  9. A. Different Anonymous (History)

    Harmonics — who cares:

    Why not just rectify to DC at each centrifuge with a suitably big capacitor and inductor to smooth out the fluctuations, followed by a …

    Voltage Regulator.

    Once can then run a DC to AC inverter if the centrifuge motor needs AC.

    We have a bunch of fairly low power motors (what, 1 to 2 kilowatts each ???) so why do we worry about regulating the power.

    Buried lines are marginally more secure from bombardment.

    Forget the harmonics.


    • Hairs (History)


      I think the harmonics cannot just be “forgotten”.

      I don’t work in the field of export controls and non-proliferation, but the people who do are, I like to think, anything but idiots – and if the export of *low-harmonic* frequency converters at speeds > 600 Hz is strongly controlled, then there is probably a very good reason for that.

      There is also likely to be a reason why Stuxnet targetted the Siemens controllers that governed the frequency converters (specifically hijacking the demand frequency whenever the controller was operational between 807 Hz and 1210 Hz).

      If the solution to powering centrifuges were as simple as you suggest then there would be no reason for the Nuclear Suppliers Group and national governments to regulate frequency converters, and there would have been no incentive for the creation of a software worm (Stuxnet) that reports suggest was an order of magnitude more advanced than basic criminality, and which likely required the effort of one or more sovereign states to develop.

      It’s one thing to demonstrate that the harmonics created by underground cables can be easily withstood / compensated by the power conversion process, and that there is negligible impact from putting the power supply underground; and if anyone demonstrates that I’ll quite happily admit that my opinions were wide of the mark. But it’s quite another to suggest that harmonics and frequency converters are irrelevant to an enrichment programme, or easily dispensed with by the use of a big LC circuit. In the latter case, I think the evidence of export controls and Stuxnet’s targetting is that frequency converters (and possibly by extension harmonics) are fundamental to centrifuge enrichment programmes.

    • kme (History)

      It is my understanding that Stuxnet wasn’t doing anything as subtle as introducing harmonics, it was ramping up and down the drive frequency itself.

    • John Schilling (History)

      Power line harmonics can be fully isolated with a simple motor-generator set, at 85-90% efficiency. Sophisticated modern power electronics can do the same job with greater efficiency, and I would assume that western enrichment facilities prefer a bit of investment in power electronics to paying 10-15% higher electric bills indefinitely. If the Syrians (Iranians, Norks, Burmese, whomever) are having harmonics issues and can’t import fancy export-controlled semiconductors, they’ve got an alternative that anyone with basic ninteenth-century power engineering techology can implement. Nor is this in any way obscure or forgotten lore, amongst the people whose job it is to deliver power to machinery.

      I do not know what the export-control people think they are accomplishing by restricting low-harmonic frequency converters. They are not in fact preventing would-be proliferators from providing copious low-harmonic AC power to their enrichment cascades, nor even forcing them to route the power lines aboveground.

      Which brings us back to the question of why Haska doesn’t seem to have the aboveground power lines an industrial facility of that scale would normally require.

    • Hairs (History)


      I perhaps wasn’t as clear as I should have been; I didn’t mean to suggest that Stuxnet was trying to introduce harmonics, rather I was demonstrating to Anon2 that the fact that Stuxnet targetted *frequency converters* (specifically the speed control of those converters) is strongly suggestive that Iran’s enrichment relies on frequency converters – and not just some crude conversion to dc followed perhaps by the use of dc motors.


      I think you’re misunderstanding the role of harmonic suppression. In the case of centrifuges it’s not about a little bit of power factor correction and loss reduction (as you suggest), it’s about two fundamental issues:

      1) Positive sequence harmonics (4,7,10, etc) serve to increase the rotational speed of a motor. Negative sequence harmonics (2,5,8,11, etc) decrease the speed. The combination causes torsional oscillations of a motor shaft (or the attached centrifuge) and if these oscillations coincide with the natural frequency then there’s resonance. At speeds > 500 Hz those resonances are going to grow mighty quickly, and considering that the Simatic S7 PLCs, which Stuxnet targetted, can’t perform bumpless control faster than 1ms it means that at 500 Hz you’ve got just a handful of control intervals to detect and react to a resonance. At 1000 Hz the resonance is growing as quickly as your equipment can bumplessly control it i.e. it becomes effectively impossible to introduce a control signal without causing as much disruption as you’re preventing.

      In variable frequency drive motors for large pumps e.g. 5 MW, or in the static frequency conversion equipment used for starting gas turbines the impact of harmonics can be designed out to a significant extent by pushing the harmonics to higher frequencies and also by relying on natural damping at the oil lubricated bearings. However…

      2) Centrifuges famously have magnetic bearings and are designed for minimal damping. Once a resonance is excited it’s difficult to see how it can be damped before becoming catastrophic within a few tens of milliseconds.

      More importantly, and unlike equipment designed to work only at 50 or 60 Hz, harmonics can’t just be pushed into a range that’s unimportant. If I don’t like the harmonics on my SFC then I can use an 18 pulse set of gated thyristors and significantly suppress anything below the 17th harmonic i.e. 850 Hz for a 50 Hz grid supply, or 1020 Hz for a 60 Hz supply. Unfortunately these typical speeds that centrifuges would approach or pass through, and pushing up to the 23rd harmonic (with a 24 pulse system) adds ever increasing complexity in timing the thyristors to avoid overlaps, residual currents and the like.

      On a 50 Hz system the positive sequence harmonics sit at 200 Hz, 350 Hz, 500 Hz, 650 Hz, 800 Hz, etc. The negative ones at 100 Hz, 250 Hz, 400 Hz, 550 Hz, 700 Hz, etc. Thus even with total harmonic distortion of just a few percent, which would be considered excellent in the power industry, every time a centrifuge starts up or shuts down, it is going to have to pass through ALL of these speeds without resonating, and when it gets to its operating speed it had better not be near any “part-integer” harmonic that was created by the power conversion electronics themselves.

      With regard to your comment that instead of a sophisticated low-harmonic frequency converter there is “…an alternative that anyone with basic ninteenth-century power engineering technology can implement. Nor is this in any way obscure or forgotten lore, amongst the people whose job it is to deliver power to machinery…” I am such a person. I’ve spent my professional life of more than twenty years working in power stations and refineries commissioning, maintaining and fault-finding the very best VFDs and SFCs that people like Siemens, GE and ABB can build – and I can assure you that in the real world the issue of power system harmonics and achieving (sufficiently) low-harmonic frequency conversion is vastly more difficult and harder to engineer than your comment implies. Even a cursory read of IEC-555 and IEEE-519 will give an indication of the complexities involved.

    • John Schilling (History)


      I do not misunderstand. I am well aware that high-frequency harmonics can rapidly destroy finely-balanced high-frequency rotating machinery. I understand that this is not a matter of lost efficiency, but of machinery being turned to scrap metal.

      I assert, and continue to assert, that the use of a proper motor-generator set, makes power line harmonics go away. Not shifted to an “unimportant” range, but eliminated. Close enough to 100% suppression as makes no difference, for all but the trivially low-frequency harmonics. The 10-15% power loss I mentioned, is the cost of making the harmonics go away in this fashion. I apologize if I was not clear about this.

      You can have the worst imagineable input power, just this side of a square wave if you like, so long as an industrial AC motor can survive it. The only thing that will make it past the flywheel, is the purest single-frequency sinusoid you could want. There is the possibility that new harmonics might be introduced on the generator side, but A: I believe anyone who can build centrifuges and centrifuge motors can build a suitably clean direct-drive generator and B: those aren’t the harmonics at issue here anyway.

      As an added bonus, you are proof against spikes or other transients in your input power, e.g. EMP. Just in case, you know, someone might have reason to try and soft-kill your industrial facility.

    • Hairs (History)


      My apologies – I agree that if you mechanically or hydraulically couple to the grid then there would be no problem with importing harmonics.


      If you’ve followed the thread above you’ll see that my original assertion about underground cables being undesirable from a harmonics point of view is wrong or, at least, can be overcome. I guess that makes a lack of obvious power supply for a large facility marginally more suspicious again.

    • Jeffrey (History)

      That’s the only way new knowledge is created. I appreciate your raising the issue at all.

  10. Anon (History)

    BTW, Syrian officials have turned down a renewed request from U.N. nuclear inspectors to visit suspected secret nuclear sites during talks in Damascus.

    & how is hiding the power lines going into a huge-ass and obviously visibly huge ass facility “an excellent sign of concealment and deception” ? Isn’t it drawing attention to the facility — which is exactly what it seems to be doing here.

    • Jeffrey (History)

      A big building that doesn’t draw much electricity might appear to be a large warehouse and escape notice.

    • Allen Thomson (History)

      I’ll take the opportunity to ask about the late BOE, which didn’t appear to have much in the way of power lines running to it. Likely it wouldn’t have required as much power as a centrifuge plant, but still water and CO2 would have to be circulated for the putative reactor, other equipment kept running, hotel services supported.