Announcement

Collapse
No announcement yet.

JNIOR DMX Control supply issues

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • JNIOR DMX Control supply issues

    Just a notice that we are unable to obtain some of the critical components for the Model 412DMX until 2023. With lead times out 52+ weeks that tells me that there is still a chance that those items won't potentially ever be available. So I need to redesign around more readily available components. This affects the isolated DMX output that replaces the AUX port on the standard Model 412. It also affects the ride-thru power supply built into the 412DMX.

    As of today there are a half dozen 412DMX left in stock and available on the store at https://integpg.com/store same as https://jnior.com/store (whichever is easiest for you to remember). After that there will be an 8 to 12 week lead time depending on how quickly I can get the replacement design done and prototyped.

    If you need to control a DMX fixture or DMX fixtures while those are out of stock, you can get the job done with a standard Model 410. While the 410 has 4 less relays it does have an AUX serial port capable of RS-485. The Models 412 and 414 are not RS-485 compatible. There is a white paper describing how it is done (in excruciating detail) . So you can do this with any existing 410 that you may have. Not that this approach is NOT an isolated output, but that is not normally a big deal if you are running a couple of standard fixtures as opposed to a full stage implementation.

    The rid-thru power supply allows the JNIOR to skate through very brief power outages (10 - 20 seconds). Short power glitches cause controllers to reset and can interrupt an ongoing process. The other JNIOR models do not have this feature. Typically I would expect that your JNIORs are powered by a reliable source anyway such as a UPS. This ride-thru supply was experimental for us. Logically it makes sense and there are some non-cinema situations where the feature is appreciated. However, if we redesign the 412DMX on the existing 412 core we can get this to back to the market faster and with valid CE certification without needing to recertify. Although I'd like to retest everything all of the time, it is just too expensive as those granted the gift of being able to issue these certifications like to get rich.

    So we work hard to keep JNIORs in stock, and this situation pains us, but it is what it is. Stupid COVID.

    If this supply interruption concerns anyone, I would like to hear about it. We will help you work around it. We are committed to making the 412DMX available as soon as we can.
    Last edited by Bruce Cloutier; 01-31-2022, 11:12 AM.

  • #2
    I'm curious as to what part has such long lead times. I used to design DMX controllers and dimmers for https://www.dovesystems.com/ .

    While at USL/QSC, I had several cases of parts going EOL. We went through several displays in the JSD series. Then the SPI flash memory used in almost everything I designed. In the case of the displays, we had to redesign the front panel PCB. For the flash chip, we found a pin compatible part that required a firmware change. In each case, I was able to get the firmware to identify which display, memory, or other part was out there and work with the part appropriately. We also included board ID resistors and board revision resistors on each PCB in the product. As board designs changed, the microcontroller would read the value of the resistor and adjust the firmware appropriately. For example, the LSS series went through a variety of color sensors, luminance sensors, and changed from analog SPL measurement to DSP-based. The latest firmware works on all those units by reading the board rev resistors. After doing design for about 50 years, I developed some tricks!

    Harold
    Last edited by Harold Hallikainen; 01-31-2022, 11:28 AM.

    Comment


    • #3
      Hey Harold...

      On the DMX side the serial daughter board used the TI ISO35DW high bandwidth RS-485 (and RS-422) transceiver which is isolated from the digital input signals. This chip also contains an H-bridge (I think) driver for a transformer providing energy for an isolated power supply. Its a nice little component but we cannot wait into 2023 before we can build anymore product.

      Stuff goes EOL and we can deal with it. In this case there are many items that are just not available and that without warning. In my mind it is mismanagement of the whole COVID situation. It doesn't matter who or what is to blame. We're just trying to stay in business and support our customers. It is very frustrating. I can rant (or whine) about this but...

      Suppliers are using magic lead times like 35 weeks or 52 weeks. It is code for "we don't really know what we are doing and when these things will be available". I can imagine that during the shutdown a lot of (very well-paid) employees of these big corporations sat home sipping reserve wine and playing video games thinking that "none of this is their problem". I read it as "time to redesign with parts from real reliable manufacturers".

      Oh, they also blame the automotive industry for their "sudden"(?) requirement for electronics. I was happy with cars the way the were actually. Who is asking for these self-driving things anyway? I'm not letting one drive me off the road. More parts to fail and maybe to kill you. We can't make a PC that won't crash (or boot in any reasonable time) and we let them drive a car?

      Hmm. See. Rant.

      Comment


      • #4
        OK, that does sound like a pretty special part. On the DMX transmit side, I used standard non-isolated RS-485 transceivers. I used a part from Maxim since it had the highest ESD rating. On the receive side, I was designing dimmers, so I used the same part, but the entire electronics was floated to avoid ground loops and common mode voltage into the receivers.

        In a letter to the editor of Electronic Engineering Times in the 1980s or 1990s, I described the electronic component market as an unstable control loop with high loop gain and long loop delay. At that time DRAM chips went through boom and bust cycles. A bunch of companies would see a shortage and build fabs which came online several years later. Each of the companies was hoping to fulfill the shortage, so we ended up with several times the required number of chips. The fabs were slowly shut down or redirected to other products resulting in another shortage.

        I think we are now seeing the "step response" of this unstable control loop.

        Harold
        (recovering from covid and doing well)

        Comment


        • #5
          Back then I was using 32KB low-power SRAM in quantity but no one gave us the opportunity to be allocated. So at the end of each month I would get calls (no cell phones) from distributors with everyone's unclaimed memory allocations and I would negotiate on the spot for the parts. This kept us afloat.

          As far as this specialized TI part, there was no reason to expect that there would be an issue with it during original design of course. At this point the safest approach would separate the digital isolation from the transceiver and also separate the isolated supply. This discrete approach leaves options for part crosses. I can get a Maxim part with the isolation but there are multiples of options for parts without even from Maxim. I am also thinking about a circuit that we can also use in an expansion module so DMX can be an option of other (existing) JNIOR models.

          That is a really smart analogy with control loops. You know that delays in a control loop are the killer and, well, to say the market (supply side) is "unstable" is being kind. Its an untunable PID loop for sure. I can understand how difficult manufacturing has to be under the conditions. It would just help even a little bit if when you talked to them they had even the slighted compassion and concern. It would be nice to know with confidence what the plan is so I can make intelligent decisions (it I am even capable of that). But the people that I can get on the phone are just happy to have their jobs and really don't know the answers either.

          I also had COVID but before March 2020. For me it was basically the mother of all chest colds. Kept me in my recliner for a week. Still went through a lot of beer. I hope you feel better soon.

          I also had an episode of Gout which kept Kevin and I from attending CES 2021. That was disappointing and I think worse than my COVID experience. It was another recliner event. Sucks getting old. I am tired of it but... the alternative isn't (yet) attractive.

          Comment


          • #6
            On the isolated driver part, I remember Analog Devices making some parts ( https://www.analog.com/en/parametricsearch/11064#/ ). I know there are a lot of mergers (other companies absorbing National, Linear Technology, Maxim, etc.), but I think Analog has been buying instead of being sold. MAYBE they have a suitable part (but probably not pin compatible).

            One other contributor to the control loop instability is buyers double or triple ordering to cover their bets in a shortage. This makes demand look like it's double or triple the actual demand, increasing the loop gain.

            My covid recovery is going well. My wife is in the middle of it right now.

            Harold

            Comment


            • #7
              Analog Devices always scared me. Talk about a company that will EOL a part. I swear, one could watch an AD part go EOL while they were still in the beta/sample phase of the part's life.

              Comment


              • #8
                Products do seem to have very short life cycles. Maybe they could save money by printing the EOL announcement on the back of the new product announcement.

                Harold

                Comment


                • #9
                  Sometimes, I wish there would be a law like... if you EOL a product within 10 or so years, you're obliged to open up the designs and/or give a free license on all the patents that it involves...

                  But it's not just the short life cycle of products that we're facing now, our global supply chain has been a mess for more than a year now... I recently ordered a new notebook, I'm happy if I get it in may. I've been waiting for a new car now for more than 12 months. We recently ordered a bunch of networking gear, due to be delivered in June... Have something break that's out of warranty and you may have some major problems at hand.

                  Not that long ago, I read they tore down a bunch of old, but still perfectly functional clean rooms around here... Maybe I should have bought them for pennies, buy a bunch of old wafer steppers and start producing all those ICs nobody seems to be able to produce and sell them at a 500% premium... Most of them don't require the latest and greatest 6 nm process, heck, most designs of what's missing right now are more than a decade old and can probably be reproduced on a 20 year old machine...

                  Comment


                  • #10
                    I think the global supply chains have highlighted a problem with how things are manufactured and distributed. Too much is being farmed out. Yeah, people like cheap goods and services. But they also like clean environments and non-slave-labor. However, it would appear that pollution and essentially slave labor is just fine, so long as one need not see it. And, if machines are making most of the stuff, why does that machine have to reside half-way around the world?

                    Comment


                    • #11
                      A high school student made chips in his garage.

                      http://sam.zeloof.xyz/second-ic/

                      Comment


                      • #12
                        Originally posted by Harold Hallikainen View Post
                        A high school student made chips in his garage.

                        http://sam.zeloof.xyz/second-ic/
                        It's obvious he DOES NOT have an every day garage shop...

                        Comment


                        • #13
                          Great discussion.

                          Most of the world in markets like cinema rely on the JNIOR. I am sure that you would be shocked to know much how we've pared ourselves down in order to survive the shutdown and to keep delivering product. Presently we have our product manufactured literally 50 feet from my desk. In fact the PnP machine just finished a new 412 PCB and it is waiting for me (yes me!) to go over and check the board, move it over to the oven and then start another. Kevin, you know him from support, is actually applying connectors and running those boards through our selective solder machine. Jake, who answers the phone and handles shipping, is running boards through Program & Test, and boxing them. Pre-COVID (in the "before-time") I had a staff handling production. But numbers were higher then. We've not recovered to those levels yet. Not close. We are doing what we need to to continue to provide the product for all of you guys.

                          As for ordering to hedge bets on shortages. I am guilty to some extent. I am sure that distributors get excited about the EAU numbers that they estimate from our orders only to wonder where that business went. I have over a year's worth of many of the critical components like the microcontrollers and relays.

                          Now I might have to tell someone that the 412DMX is Out of Stock and that pains me.

                          I bought my first real computer in 1979 for around $10,000. Now every couple of years we throw away a computing device, your phone, that has 10,000 times the computing capacity as my first system. And, now they are trying to make cars that go obsolete/EOL every 2 years. No thanks.





                          Comment


                          • #14
                            I use on our automation the ST232C and it is getting long lead times as are many of the Pic chips yesterday I went to order some trim pots and they were 52 weeks delivery

                            Comment


                            • #15
                              Well, the new Intel campus being built in Ohio is nice and all, but I doubt they make any of the chips you fellows are using.

                              Comment

                              Working...
                              X