Supply chains are really messed up post Covid and the 🇺🇦 war.
New supply chains are needed to give independence from the pre Covid global supply chain model. Countries need to establish new sources and new manufacturing capabilities. The new supply chain models will have to have multi sourcing to address specific future shortages. The supply chain will need new government leadership, trade agreements and country level independence for national critical infrastructure.
Mercedes/BMW... all the electrical harnesses were manufactured in the Ukraine 🇺🇦. EU is now on 20 month delivery delay.
Microchip ... current chip shortage is unlikely to be resolved in the near future, partly because of the complexities of the semiconductor production process. Typical lead times can exceed four months for products that are already well established in a manufacturing line. Increasing capacity by moving a product to another manufacturing site usually adds another six months. Switching to a different manufacturer typically adds another year or more because the chip’s design requires alterations to match the specific manufacturing processes of the new partner. And some chips can contain manufacturer-specific intellectual property that may require alterations or licensing.
Innovation in the industry is needed. I voted for a focus on the new chip types with legacy chips being of secondary importance. Innovation is needed to take the old code base and automatically convert to the new chips - this is needed more than ever. Microchip should focus on innovative methods for code conversion not muck about with the options in the MPLAB MMC. It is not hard to convert old code. What is hard is finding the engineer with the experience to make this happen.. the odd intern may be able to help but not lead the proces.
So, in essence. I wrote a white paper in April 2019 on this subject. It was not hard to see the issues with the globalisation of supply chains. It just happened as expected.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
We have a large, large number of products out in the field that use the 16C54/55/57 range of devices. Switching those to a newer chip would be a massive problem for us. We don't have the original source for the code. A few have some (very old, now deprecated) .asm "source" that I can't understand. Almost as bad is that we have no circuit diagrams for the products. The person who designed these things has left no documentation of any sort whatsoever. I might be able to sit and re-write everything, for all of the five devices in our product given the time. Although even if I could decode how I think it is supposed to work, attempt to replicate that, and make it work, the devices that might be suitable are in equally short supply.
I designed a power supply a few years back which uses two STEF01 devices. These are not available and we needed some making. I tried to redesign it using similar devices, found a few that looked like they would work. No stock. Checked Farnell for stock of some small PIC processors and found that they had six or seven thousand of which ever one I thought I could use. Completely redesigned the PSU to accept the PIC in question. Ordered some PCBs. Went to order the PICs so I'd be able to get soldering when the PCBs were delivered. No stock. In under ten days those six thousand devices had gone. Built the five prototypes using some stock I'd got. Had to redesign the board again to take another device where the A/D inputs are on different legs. Ordered more PCBs and more chips. Assuming the third revision works when I make one up we'll at least be able to make the fifty that we've got chips for. After that who knows? A pity as the original with the STEF01 has been rock solid.
As well as making our own units, we repair other equipment. One of these uses a motor drive chip that has been discontinued by Infineon. It is on old chip, granted but nothing else fits on the all surface mount PCB. Enquired with a supplier who specialises in obtaining obselete components. I asked for fifty devices. Supplier came back with a quote of seventy five pounds. I said, "that's not bad, they're not much more than we were paying at about 80p each". Supplier replied: "No. they will cost seventy five pounds EACH!". Attempted to order some from China and we got a price of one hundred and fifty pounds each, shipping not included.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Lack of Capacity Planning ? :>)
Supply chains are really messed up post Covid and the 🇺🇦 war.
New supply chains are needed to give independence from the pre Covid global supply chain model. Countries need to establish new sources and new manufacturing capabilities. The new supply chain models will have to have multi sourcing to address specific future shortages. The supply chain will need new government leadership, trade agreements and country level independence for national critical infrastructure.
Mercedes/BMW... all the electrical harnesses were manufactured in the Ukraine 🇺🇦. EU is now on 20 month delivery delay.
Microchip ... current chip shortage is unlikely to be resolved in the near future, partly because of the complexities of the semiconductor production process. Typical lead times can exceed four months for products that are already well established in a manufacturing line. Increasing capacity by moving a product to another manufacturing site usually adds another six months. Switching to a different manufacturer typically adds another year or more because the chip’s design requires alterations to match the specific manufacturing processes of the new partner. And some chips can contain manufacturer-specific intellectual property that may require alterations or licensing.
Innovation in the industry is needed. I voted for a focus on the new chip types with legacy chips being of secondary importance. Innovation is needed to take the old code base and automatically convert to the new chips - this is needed more than ever. Microchip should focus on innovative methods for code conversion not muck about with the options in the MPLAB MMC. It is not hard to convert old code. What is hard is finding the engineer with the experience to make this happen.. the odd intern may be able to help but not lead the proces.
So, in essence. I wrote a white paper in April 2019 on this subject. It was not hard to see the issues with the globalisation of supply chains. It just happened as expected.
We have a large, large number of products out in the field that use the 16C54/55/57 range of devices. Switching those to a newer chip would be a massive problem for us. We don't have the original source for the code. A few have some (very old, now deprecated) .asm "source" that I can't understand. Almost as bad is that we have no circuit diagrams for the products. The person who designed these things has left no documentation of any sort whatsoever. I might be able to sit and re-write everything, for all of the five devices in our product given the time. Although even if I could decode how I think it is supposed to work, attempt to replicate that, and make it work, the devices that might be suitable are in equally short supply.
I designed a power supply a few years back which uses two STEF01 devices. These are not available and we needed some making. I tried to redesign it using similar devices, found a few that looked like they would work. No stock. Checked Farnell for stock of some small PIC processors and found that they had six or seven thousand of which ever one I thought I could use. Completely redesigned the PSU to accept the PIC in question. Ordered some PCBs. Went to order the PICs so I'd be able to get soldering when the PCBs were delivered. No stock. In under ten days those six thousand devices had gone. Built the five prototypes using some stock I'd got. Had to redesign the board again to take another device where the A/D inputs are on different legs. Ordered more PCBs and more chips. Assuming the third revision works when I make one up we'll at least be able to make the fifty that we've got chips for. After that who knows? A pity as the original with the STEF01 has been rock solid.
As well as making our own units, we repair other equipment. One of these uses a motor drive chip that has been discontinued by Infineon. It is on old chip, granted but nothing else fits on the all surface mount PCB. Enquired with a supplier who specialises in obtaining obselete components. I asked for fifty devices. Supplier came back with a quote of seventy five pounds. I said, "that's not bad, they're not much more than we were paying at about 80p each". Supplier replied: "No. they will cost seventy five pounds EACH!". Attempted to order some from China and we got a price of one hundred and fifty pounds each, shipping not included.