I'd Love to know the design phases of this thing. You worked on this? Yeah, yep. well, you designed the front panel you work intimately on managing. Did you manage the development? No, no, no.

Okay, so I'm a technology manager. One of the things I do is I work with teams to create common building blocks that get shoes between all the projects I've got and I'm often brought in to help them go over a particular difficult technical challenge. So for example, on the high end function generator, they had some horrendous problems with the switching power supplies to power up all the high power Asics creating Spurs yeah yeah into the output of the function because the 120 mega oh yeah, it doesn't take much Dbid noise output and I brought in some power supply people to help them redesign all the power supplies to reduce the noise because our power supply group is really good at low noise switching. they were not good at low noise switching, but they had tons of power requirements cuz they're giant, a giant Fpg I should say and some really high current DAX and eight of these they're not DVDs but you know, just high power circuits.

So the the switching circuits they put in were like larger than they had ever designed before and they're not experts on switching pallets. The folks in New Jersey are so I got involved and I got another guy involved and we redesigned their entire power subsystem for them so they could stay on track on the project. There's when they hit a wall rather and say you guys better fix it and work 12-hour days for the next three months. No, that's not how we're gonna do it.

We're gonna bring in people who know how to do it and get it right quickly so you don't delay the project three months and don't burn out. We get it done right and get it done once. So I'll do that I'll The folks who developed RSM use didn't know how to do this kind of graphics interface so I had the the team that developed the the firmware for this and our hardware design for this and we transferred that stuff to our Japanese team that did the SMU So the mountain the Mount Fuji project was our new SMU to compete with her folks in Cleveland and they leveraged all that stuff from the other designs. They use the same same circuit designs.

They took the schematics and transfer them and then they built it up from there. That's what I do is I help make those connections. So the development process. you need a replacement.

Mm-hmm for the buddies at 3400 401 34401a Juke Um up was it. What shipped Was that the original concept? How many ever little? wasn't a evolution? did you go through? Did you go like we're gonna have a graph Alou screen. It's gonna have the it's gonna have the graphene and it's gonna have this and how I already had the function generator at that point. Oh my captioner was pronounced the first major product at the counter and the function ever developed more than us in parallel.

So we had the basic structure but it didn't meet the cost target so that's when we had hi and crush it and reduce the cost dramatically and that's where I got heavily involved. But the DMM technology believer not was actually taking the we had done a research project on replacing the 3458 which we ended up putting aside and that team stopped working on that and then we had them go and do him dramatic cost reduction to fit it in these kind of price point interesting and you strip it down and simplify it. and then the team you know in Loveland Colorado site came up with the devote meter to make it meet the cost targets. I came up with a new low-cost front panel and we merged the whole thing together.

but this architecture is dramatically different than the 34401a s the 3441 that the fundamental problem with the volt meter is that it's isolated and the 34401a the entire processing core next to the measurement engine. Yeah, and it directly controls the measurement engine and the entire front panel is floating right next to at the main potential. that worked back in the day of GPIB and Rs-232 because we had isolate GPIB and Rs-232 over very simple you arts as a chip called the mouth protocol which basically takes and isolates that and brings it across isolation into a custom ASIC and that's great. for GPIB and Rs-232 by the USB and LAN, it can't be done.

Yep, you need processing right next to the LAN and USB because it's all grounded you can't isolate must be in land very easily at the high speeds. So you change the architecture entirely to put the processing grant which to control the screen and did all the processing right at the back end. so to speak that's correct that that we're all the ground potential. Yep, so the front panel display also because this is would be floating.

that was a problem. Isolating this what's safe with a flat panel display is a little bit more challenging than with a vacuum fluorescent display. Ah, a little bit easier to keep the cracks from giving you you know shock hazard. So yeah, this also the LCD display has to be directly connected to the processor.

Yes, all right 24 lines I mean it's a very wide bus so all that had to get moved to ground. so they had to develop a whole new protocol to a much smaller set of hardware that controlled the actual measurement engine. Yes, they developed a whole new system to do that for this project and that's what enabled these products to do what they did using the same processor platform for the most part that's in the counter, the function generator, and our power products as well. Karna Sowwy.

They ended up using FPGA for the measurement. There's an EPG and a single chip and a single. What's the single chip? The single chip does the communications across isolation. yep, does the slower state machine.

If then the FPGA does the high-speed processing of the Delta-sigma A to D. So it does the big fir filter that basically takes this this A to D converter. No doing high-speed sampling the multiple bit and then turning that into a slow, very high-resolution reading. So cost was paramount.

Was it the entire driving factor behind this said? well, there were two. One would be complete compatibility with the existing product because it was so ubiquitous are divided into intestine by Mississippi Largest selling product in tester matters, your company is like a stream, probably in any other tested measurement test images to we're the big but the 34401a tits peak sold over 25,000 Yeah, we're not back up to that with this, but we head in there with some of the disk. Competitors have come in and taken some way. but and it's growing.

This is just the rolling quite nicely. But uh so there was compatibility and there was cost. Seemed to be the - absolutely driving factors nothing yeah well and also differentiation. We have a have nuclei abilities we have do the new graphic system yeah yeah, we've added data logging.

We did histograms things like that which you know today. they're very useful exactly. but the point is you got a graphic explain. damnit.

Use it. Yeah, you know, Yeah, and that's it's. not just a different way to control the thing because the menuing system on the 34 for one is not as easy to use. Nice.

but it's hard to justify switching all that just for that, right? So you might get more data and there's more things we'd like to do in the display. But you know. and they did some newer things. I Think with the 65 and 70.

if not mistaken, they have dual parameter information. Yes, Yes, yes, I Do there is a Judas light right that wasn't in the first one. We didn't have enough time to fit that in right they wanted to. But when we got to 65 and 70 day today and frankly, with a 34 for 10 and 11, they had the dual parameter in a vacuum fluorescent and when we first introduced the the more basic models, we didn't have it and we got feedback, that was a problem.

They liked the dual display all right, so we fixed it and they rolled it up into 65 and 70. The biggest challenge in this was when they developed the 34401a early. They were on the second floor of a building and the manufacturing center was on the floor. Good.

So it was very easy to optimize cost and learn what how you're gonna cost us with the rules for engagement work to optimize in the right way. Now this thing is manufactured in SEM not even interpreting operation Oh like I had our the supply chain of the Transformers made in Kuantan Malaysia right? Okay and back then we bought it from a local vendor all right and therefore it optimizing it the perfect shield India at the common mode noise all those things was much easier. You've got in your car, you drove to their site in an hour, you have meaning you get the thing, then you gotta move the shield this way you know, no problem, he's now it's a third party through our penetration back to the States and yeah, and getting the cost of it and getting the performance of it is much more difficult. So one of the other things I do is I help I go to Penang fairly frequently and they had an issue with the vendor we were using.

they weren't meaning the cost I actually traveled to China to visit another new company too because I do a lot of magnetics materials work for people to power supply background and I bet it visited the News Cup this new company and vetted them out and made sure they could do a quality job and and help communicate our technical needs and they're now meeting cost on that. Much better than they were before is the reason you went to a contract manufacturer Cm for though you were using the terminology there is it. is that purely a price point thing or a bit like why don't you mean you've got your own huge manufacturing facilities? We do. It's an interesting story, some of it.

I think in retrospect we might not have done, but let me play the whole thing out. First of all, keeping up with the latest surface-mount board loading. It's very expensive and it's constant changing by. You're not going high density and stuff like this.

We're not as high density as some products, but we're certainly using smaller and smaller parts everywhere and find pimp each BJ's and I know for oh twos and even up to Oh ones in some cases. And right, D offends which there are almost as bad as BGA's yeah and you know we're moving toward you know, three and four thousands traces and spaces. Oh okay, things get denser only because that's what the packages are coming out and they're smaller and you try to fit more functionality and less space you have to. But the point was our company made a decision.

get out of board manufacturing cuz we this was back. We got back out after just went surface mount and equipment because already starting to evolve and change and keeping up with it and doing it in a quality way and as Rojas happened to yeah we didn't want to make that investment. Leave it to the you could buy experts. yeah you combine that service.

So we moved the boards out and when we moved the boards out we still did the assembly of the final product in our own site, but then we were still part of the. this is kind of a related thing is that we were still part of Colyn at the time, Agilent was still building their life sciences products in the States in fact, in Delaware and a few other locations and they looked at our cost of manufacturing much lower in Oh. Neither about to invite. Well, our factories only so big.

Yeah! and they said you guys got to move out to make space for them to move in Oh So we basically started. We went to the CMS we're building our boards and said we want you to build our products and they said there's no problem I have your business And yeah, we started to move a lot of the products out of our local and the idea to us that was to lower our indirect overhead. They were more efficient at manufacturing because they do a whole lot more of it than we do and it was even less expensive. And also one of the other benefits is that they often have small shops in the US that we can work with for prototypes.

even at the final product level. And for some products where the volume is are very small, we would just stay in the US Yeah, we could. We could do more localized manufacturing and get quicker turnaround and faster response time for low volume specials and high-value products. So we made a decision to move our manufacturing out of our factories to local vendors right in the Penang area north down in Johor and places like that.

That created some challenges. some trained challenges. We're not completely through that, but we're mostly through most of the stuff that's moved out. some of the very old products we did MOOCs We're not going to keep them for that much longer, so if you go in in Penang now you'll see production lines for our products.

but they're mostly the old through-hole old. Interesting that we don't want to spend the money to move because it costs a fair amount of money to transfer all the test sets and training and all that kind of stuff to other people. So one of the challenges for this was this was introduced as we were starting to make some of those transitions right and we were still able to keep. We tried to do the new products in our factory and then move it at first.

Now we do direct to CM. The new products go right to the CM first product. This product was still built in Penang operation. They moved out after it went into production, but the newer versions of the product, they're going right into that I Noticed that I've got one of the original made in Penang once.

Yeah, but everything's made in Malaysia Yeah, it's toughened made of our factory, it's It's an interesting challenge and I think the Cioms have been exceptional. Frankly at doing a good job. it still takes a lot of scrutiny on our part in review and keeping the quality high. but these are big big corporate companies.

They know how to do a good job and they we think the flexibility long term is going to be good for us. Thank you! Able to move things around. You know customer needs are special and they have a facility one of the three I'm not going to name all that seems we use, but he's really good at transferring between the different sites there. That's one of their skills and we're seeing that as a potential way for us to respond quickly for modifications and specials and needs.

so we'll see how that pans out. Some of these things you know I wouldn't do myself. mm-hmm I'm not saying this one is risen, but if this one's worked out okay and our goal is long term is to lower cost and get more responsiveness back, do you know what the issue there was? Absolutely yeah. you do tell us Very interesting.

Stella Oh Tell us the interesting story about how this was Respect. It came out of the Thousand Volts spec and then did somebody realize? oh, did you guys realize that there was an issue There wasn't an issue of, it wasn't an issue I said he's blown out of proportion. Did it blown out Okay, do-do tell us. Anyway, So what happened was we had a customer who was very interested in our voltmeter to reduce the amount of calibration they had to do and they wanted us to create a custom version of it they've built into their system.

Oh, and you would do that for a big enough customer, for a big enough customer that would pay us enough, Sure right. And it was also strategic in that we wanted to learn more about their needs. So sometimes we justify things by saying this is a door into learn more about their needs and in certain growth areas. We'd like to do that.

and the team that was gonna do that was in one location and they took the design of this and they were gonna reform factor it into a card that went in the customer system. So they're going over the design effectively doing a complete design review and they notice that one of the parts in the design the the voltage across the part was above the rating on the part. which pop do you know is it not gonna say okay and what happened was they said this is unacceptable and we can't do this and they they said it's a safety issue which it wasn't and they alerted some people and and panic ensued and panic ensued and the knee-jerk reaction was to make a change to safety. First that pie change to the spec into the spec and don't do that Because yeah, we said this could be a safety issue and the reality was that our further review it was not a safety issue.

But but you did publish the spec. you did revise it and published it and hence why it was noticed by people on the forum or big time. Sure, Yeah, yeah, it's not a safety issue right? And the part is accepted to use in that application that way. And since then it pushed it back to a thousand volts.

And so you've changed the spec back and everything's hunky-dory Yes! so there was no changes made, no changes made new, people looked at it and said this isn't right, right and it turns out it was right. But people is one of these things. We've done it that way for so long no one remembered anything about, but it turned out one of the things that got it back to the right was we took our competitors products. they did the same way.

And finally the people who had not looked at this in a long time said you know, maybe we need to rethink is this really wrong and figured out it wasn't what's okay and it's back on track. You know we don't get everything perfect every time. Oh, there's such a complex set of situations. safety is we take very very very seriously.

So I say if we're gonna have any jerk reaction I'll take it there, right? Okay, okay and unfortunately did upset people. It was a mistake and we undid it. Yep and everyone's happy with it now. Everyone accepts your explanation.

Well, maybe they were I Don't know if they're looking at this, but that's that is what happened in It was A. It was unfortunate that it caused the disruption in ways that our customers time wasted our internal time. But the fact is, you get judged by when you get it right? Yep. Dan Yes, we're sorry that we accidentally thought something wasn't as safe as it really was.

Yep, but that's not on the cautious side. That's right. Yeah, a lot of companies would have. just don't worry about that.

We probably could have gotten this one right sooner and not have the option you? Okay, I think it was a little knee jerk on our part, but if we're gonna do a knee jerk, I'll take it here. Yeah, okay, so that's what interesting I was actually involved in that. Oh, I can't because our supplies have a lot of safety issues, especially switching off line power supply. so I was pretty familiar with it and I was one of the people that knew it wasn't a problem, but the certain people thought there were too many opinions in the room and told me to stand down and I say right? Okay, okay and finally came around.

You know I'm not saying I'm that I was right I'm glad they got it right. That's what matters. It wasn't one of our best days. look honest about it.

All worked out well. Yes. Excellent site. um compliance.

You know the cat testing and stuff like that, Not so much the bench meters but hand very important in handheld meters and things like that. Have you ever sent something away? that's you know, like failed and they've come back and say hey or hey, this is not gonna pass or something like that Or you do. You guys always get it right. First time, so excited it right the first time.

Not it's a challenge and frankly the new IEC 61 oh one yes you comply with it's much more challenging. Yes, one of the subtle things. What's what's the major change for those one of the biggest changes is that you have to anticipate likely mistakes the customer might make. All right, that course the safety hazard.

It's a problem as in operational mistakes. Probes in the wrong, Everything frozen there is. So for example on our new power analyzer which can measure a thousand volts all right and isolated 50 amps of current ice laughs and a thousand volts. we have BNC s for the the current input if you want to put an external transducer on the high input and we went to put the triggers that are at ground reference we can see because they could plug the beam on today's show.

Two different connectors? Yeah, no, you can just put a label on the back saying don't do that Robin I Anticipate likely mistakes and it still has to be safe. We could have had a hook up to there if we made it a 25 amp dancer. That's the kind of thing I Know that's gonna start shifting us into yahweh's you can't just put a label saying if you're stupid, you're gonna get hurt. It's not accept it won't pass compliance.

That's correct. It's relatively new and but that's not a bad thing. No, it's offensive. It's a smart thing, but you know certainly something that you see examples where that was the that was the way you handled it.

you just said, you know Oh can't we just put a label? You can't You don't do that right. You can't just put a label anymore. Yeah, we're having an issue in some certain the power supplies where the the binding post can only handle 20 amps. All right and we're gonna.

We're gonna have 40 amps so you know what do you do? Well, they pull 40 amps out of a 20 amp Bunny post. It melts. yep and it can create a hazard. It's hot to touch my bridge games so we can't do that anymore.

Used to just you know we put a sense leads on the front panel that's a you know 20 EPS max. Oh and then it throttles it. No no there's no throttle, it just says don't do that Oh hold more The design is like that's not acceptable anymore, huh? It was a purple I see the new the Referee of the IEC 61 I won. Oh so those are some of the standard challenges and one of the other things we did many, many years ago for our power products was we switched from an offline system to a 48 volt grounded system.

We use a distributed DC system and that's made safety a lot easier to meet for us. shorten development times because even though our output floats a lot off grounds with a 240, our input is only 48 volts and it's grounded on it as opposed to being 400 plus with a PFC front end. So that's actually simplified safety for us a lot in shortened development times and made our performance higher and not directly because we found that the time was taken to all that primary related circuitry. The past safety was was increasing test times and compliance testing and certification dramatically.

Now we buy a pre certified brick you know, but handles all that thing and you're seeing that with a lot other products I use. They brought these with choke snake things or they you know, the phone charger type solutions for low power and all the safeties inside that? Yes, sure that's right and that makes things a lot simpler. You're going to see more of that for low cost. You can see it for a low-cost Voltmeter Oh Interesting.

Okay, because it's just to do all the compliance testing. And yep, the cost to do it is. It's quite a bit higher. And those choke snake things and the little plugs in the wall for the phone charges are getting higher and higher in power.

Yep, you know there nowadays you can watch crazy. Your phone your phone can put out can take in like 60 watts. Yes, it's mask for a quick charger. You can take it in like 12 volts at 5 amps or something like that.

So you can get little tiny plugs that can put out a hundred watts. You know it's one unit and that's more than enough to power a Voltmeter. Oh, it's crazy. Speaking of power is just standby power consumption.

You mentioned power factor correction. All that sort of stuff. Is that a big deal in test and measurement. Is it a requirement? Like do you got a requirement? It's not.

It is something we care about I know the new one we just came out with. People are a little upset because the the power when it's off is not really off and we're actually working on the newer ones and we're gonna make it lower then you will want which product I Can't tell you. Oh oh this is an internal still under development. Ah rock I need to refine that and somebody internally went hi this is drew an x amount of water.

well they don't The fan was spinning very slowly. Oh right, you know it's the famous spinning because we don't nobody turn power off and it turns out that we use a PWM fan and PWM fans can make the fan go much slower so it's quite, but they never stopped. Completely those new designs. We have to get a switch to turn the fan off so you won't have notices.

Yes, it's not. so just a little nuisance if you will so there's no requirement. If a scope tight you know has a power factor of 0.5 or something you don't know nowadays that once you power five full power PS fees you know even low power, the PFCs are available down to what less than 100 watts. So you will have a power factor.

Almost any product today will have a power factor of 0.98 0.99 All right it used to be PFC was only used at 600 watts and up now it's even a 45. What Brick has, right? Yeah, Because the chips have gotten so cheap and it's actually above a certain power level, it's actually cheaper to have PFC than none. Ah, the components, the amount of lost savings and the parts makes up for the extra cost of a circuit. I'm sure see the whole PFC fits in a single little hybrid that they make for 50 Cent's Oh really yeah, it's Wow That's why you still need to choke and think yeah, a little bit at your input filter.

but uh no, we we do care about power efficiency at full power. I Thought you were asking about that pice. Hey Chief, a question. We are interested in getting the power.

They were on the bench. People do not like when it sits there and consumes a lot of power. Yeah, but there's kind of a threshold about 10 watts energy stars and a 10 watt number in the U.s. that means equivalent in Europe that they like it to be below something.

but it's not regulatory for test and measurement equipment. it is from Zoomer Products is the law for consumer. God and God Yep, right? So Test and Measurement gets a free ride. Well PSC We got a free ride.

but we start attention to it because yes, it's a free ride for us to get. Yay! All right, but we do. We do have to pay attention to it for standby power because people don't like that. those big ugly push switches Oh What do you mean big ugly I like them.

mmm big clunk and pass but it takes a lot of space up. Yeah I know yeah you gotta have a plastic rod to go back. Break: You wouldn't believe how easy those things break. Yes, yes they do.

I've broken a few doing today. Yeah, well, not even know it's break during a some cheer and assembly. Ah right. Oh maybe then get a bit brittle and then break in the field.

That's right Hi So everything it so we do towards we're going to more soft switching something. It also helps boot times because we pre boom guys it's it's sort of waiting. it's halfway through the boot process so you hit the switch. it comes right back up right.

but it does draw slight amount of power and we're working to get that lower and lower over time. How low can you get like what they do you have like a tile like anything under a walk because it's a nice round number. or well we're like limits. It is because we're using bricks applies here on this is using a transformer which yes the Transformers engaged and the the power supplies on the output of that are engaged.

So this is actually there's a limit to how low we can get this but would be the little supply the brick switching supplies. They actually are designed for consumer products so they can often spec they go less than a lot when you haven't standby and then they have a small power supply that comes out that keeps a little bit of circuitry alive. But why on your TV yeah has this a small amount of circuit? Then they infrared the infrared to detect you want to turn it on Yeah so that that architecture is slowly seeping into our products to minimize the power and you can get those down to like very very low. like Condor milliwatts right? As to that, one button is separate from all the other buttons here.

Yeah, speaking of the competitors, Yes, like especially in the power supply market I Mean it's a ridiculous what you can get for like on you know you can get a power supply for 50 bucks on eBay mm-hm including delivery. Probably right? No, maybe not about that cuz there it's still the box. It's just crazy. Are you at what price point? at what price point of instruments will you not go on? Is there a price point where you'll go look? we won't bother today.

There's probably a price point that's 500 I see the new scope? Yep, but that's just a temporary situation. Our low ankle long-term goal is to go long, go where we need to go to be effective interest. If someone makes something that we don't think customers should have on their bench because it's gonna cause more problems than it's worth, right? We won't go there obviously. just cuz someone else makes it, doesn't mean we want to put that product out so that $25.00 product may be fine if you're powering up a non-electronic for example, but it's gonna have tons of ripple.

It's gonna be a slow. it's not gonna necessary. be reliable. Yeah, be easy to use.

You know we're never gonna make that product. Yeah, but there's There's no reason we can't do a 500 our product as evidenced by YESCO Of course we can. We can address those spaces right. Empower in everywhere.

It takes time for us to turn our organization around to really get the efficiencies and the infrastructure we need to be effective, but we're on a path to do it. Awesome. Is there enough margin still up that low-end there is to make it? Or do you have to make it up in volume? Like well, you? obviously? Volume by volume is the point and Vine makes a big difference. Common design, common parts, and Mmm.

excellent engineering. Got it. It really requires some of the best engineering. It's actually to me just like you're doing a six and a half or even in a tin.

have DMM requires some of the best engineering in the world. Low cost product design is an art unto itself. Component obsolescence. You clean in about meeting the but its total wasted effort.

You know you already got a design that works. You can ship it and you know, spend money that make it continue to ship. And the problem is anything you change in that digital domain, you have to repeat all the testing. All the environmental testing are five years ISD because the new chip could be sensitive when the old ship wasn't Of course, that's of course.

big bucks. Yeah, and strangely enough, you may have tested it last time and gotten lucky. Oh, I didn't know. And then you test.

My second time and had nothing to do with the changing God. Yes, because last time maybe they change the test setup slightly or the unit's you test last time a little less sensitive and you fail and had nothing. so you end up fixing and something it wasn't what you started out to do. Wow I Suppose common parts, they go obsolete.

Ramen Ramen Ramen Ramen before because De Di has registers that you're gonna write specific. DDR has registered. you have to configure. so you have to change the code Sometimes right? Change the Gr right? Yeah, because I want to thought dollars just pin-compatible right? They sorry jelly.

Sometimes it is when you start to do all the testing right and then sometimes it is and you have to change the configuration for the DDR. That's one of the once we run into the another big one that's been a big headache recently was and I think it's more of a glitch in time was FPGAs Alright yeah, one of the two big FPGA vendors moved their foundry from one company in Taiwan to another company in Taiwan Yeah, the first company I guess wasn't working for very well wasn't on the latest the greatest technology and they decided they didn't want to build anything at the old company. So they started obsoleting a lot of the parts that we're making an old company. Well, it's extremely painful.

9i um they older FPGAs weren't done with full HD le where schematic base now new tools don't want that except the schematics and the documentation on that Pj is never good enough to you. Just hand it to someone else. you give me an original engineer involved if they're still around and it's just we've spent a lot of money in time doing it. We're getting good at it.

But do you have to get I've worked in in industry where we have to get like a guaranteed written guarantee signed by the CEO that they will keep this part alive for 10 15 20 years? Okay, we buy large quantities when they pass through. Oh I Can't lend you a last point. We do a bridge bar. yeah, probably is our products.

you know the test and measurement history and power supplies in particular have the longest product line. Yep, yeah, of course it 20 30 years. It's right. Yeah, now the newer product families for this particular FPGA vendor.

Since they've moved to the new foundry, they're moving out to deliver the lifetime date. They're gonna make it forward because it's working very well for them and they're not trying to get out of it. but you know, once bitten, Twice shy. Yeah! so we're taking a more full documentation approach.

We're doing more training so people can take over and move move the design more easily. In the future, it's moving completely away from schematic based designs. yeah, as much as possible to feel full HDL which read: compiles in any of the tools. All right, So that's gonna help us.

but hopefully this was a one-time glitch. The analog stuff. There has been some interesting parts that have gone away, especially in like a power supply control. You know, regulators and stuff like that have going away because companies do it for consumer products.

Like you know. they've done a really high volume part for a consumer product and then you make it better. Probably they don't sell anymore. The old ones.

They don't want to make it anymore. We designed it in. foolishly. Yes, we have to fix it.

Those are not as big of an issue. We still have to retest the whole product. Yep, you know it's a bit of a pain. Of course it's bad if they're bigger, but most of the time the switching chips are smaller the next time.

All right, so it's easier and they're more efficient. And they work better and so there's advantages to do. I Sometimes you got to remember an event is doing it in the sense that the products shipping the way it was. We didn't need to change your arthritis been one size.

Yeah, yeah. The fact that it's a little bit smaller. Well, the space was there for the bigger part. It doesn't help us, right? Certainly in a new design will move to the newer chip, but we're forced to get rid of it in an old design.

It kind of stinks. Frankly, it's it's A problem that's never gonna go away. Really, is it? It goes away by doing more modular ization. Okay, and when you do modernization and you put things in the right groupings, We've made some relatively foolish mistakes in some of our products.

We group things that should have been grouped together. So for example, this product was a good example. This product, the ADA, DS and Das are on the same board with the power supply. So that means we have five different versions of this boards.

So when that A to D changes they arrive. We gotta do. We have to babble. Yeah, so we've done in the newer products.

It's all. The digital control is on a little plug-in board, a memory module right and that snaps in and we actually use that to upgrade to performance the product. We have different versions, but also when that part goes obsolete, that board is common to all the family. Got it? Okay, and that's helping us.

We have to do it once. and then we just pop it in and do the retest. So making some smart decision about partitioning helps to the part of obsolescence quite a bit. but sometimes it's just bad.

you.