Belktronix Blog
the batteries and adding the support hardware back on the cart support rail (current sensor, power supply, etc) I've
got 2 controllers calibrated on the lab pack. What a pain that all was to get redone. I contacted the rep for the
batteries and explained what happened, but was offered no remedy for the problem, so hopefully having them
mounted normally the problem will be gone. One more controller to calibrate, and its off to the dyno for some more
work. I hope I can get by with the existing pack for the duration of the tests. After all this work on the lab pack / cart, I'm
sort of burned out for now. Time for some shuteye.   
I created this page with a dual purpose, one of which was to remind me what my day to day duties consist of as a
hardware manufacturer. The other was to give a glimpse to the viewer what's going on here as well. I hope to update it
at least weekly.
Where's the rest of the blog?
Right here
Thoughts? Ideas? email me here
8/17 Had a chassis mod to do since I had moved some components on the heatsink (a whopping 0.2"). Once that was
documented, I got the first of 3 800A controllers on the dyno, but the pack had self discharged over the last week and
needed recharging. I don't have a good feeling about this, but I'll see how it works under load. Thanks to the
customers on the receiving end for your patience. Its been a rough road to keep things progressing! Back to work!
8/19 Seems to be a challenge to get to 800A again. The highest I got to today (before I had to retreat from the heat)
was 700A before the generator surrendered its grip to the Warp 9's torque command. I've got some digging into this to
do. It may be a thermal problem with the micro shifting the frequency, so I have to check into that effect. Still plugging
away, things are nearly done with the 3 controller build. Back to work!
8/22 Finally got to my 800A goal on the dyno. I ended up having to change the current loop response on the
generator to get things more stable at high current. Once I get back from a day trip to Reno tomorrow, I can get this
one packed up and get the rest of the controllers tested and start delivering some hardware! Back to work!
8/24 Back from Reno and have one controller done on the dyno. Getting the next one ready to test tomorrow AM. Its
nice to get things moving along instead of struggling with test hardware. Back to work!
8/27 2 controllers dyno'd and 1 to go. Was tripping the Electronic Circuit Breaker (ECB) circuit which was set to 450A. I
found the AC ripple current from the generator was adding to the current that I was pulling from the pack. I increased
your web browser when visiting web pages here to get the latest updates. Its getting late, more to do tomorrow!
8/30/09  I added a 5 x 3900uF cap bank to the output of the AC generator hardware. So with that addition, I decided to
run the last 800A controller on the dyno. The ripple current coming back from the AC generator is now very low and
the controller passed with no problems. Have a day trip to Reno planned for Reno tomorrow, so more to do when I get
back. Time to get some rest for the trip.
9/2 Began the charger build for the remaining half kits that have been sent out. Have a few additional chargers to
during this part of the schedule. I'm visiting some old on-grid hardware, I need to deplete the pack so I can recharge it.
Instead of dumping the pack through a power resistor, I'd rather supplement my electric bill if possible. I don't plan to
make it a product, so it will be a hack of a few existing circuits to make it work as quickly as possible. I usually save the
fun stuff for the end of the day anyway, just in case there's a oops, I can shut down the lab and get some sleep!
Click refresh to be sure you're getting the latest scoop
9/13 Its been a busy week, while things have been moving along, there were a few bumps in the road: 1 kit shipment
had damage to a contactor (replaced), I forgot to send out the rubber mats with the kits (sent out), had to cut a 20' x
3" dia drain path to prep for the rainy season (almost done), the breaker panel in our pumphouse burned up a
breaker and power contact (replaced), and installed a Internet connection for the lab PC so I don't have to constantly
go in and out of the house to get online or check mail (done). I managed to get 1.7kW from my existing 1.2kW charger
design (nothing to do with the charger build, just curious where the limit was). PCB fabs have been nice and consistent
so thats good, I managed to load a LinkPro HV adapter that worked to 400V, to be offered soon. It has a Aux input
(non-isolated) that can be used to measure the 7W DC-DC output into the meter. Heatsinks are drilled and prepared
along with the charger cases and fan/mounting brackets. Need to update the customer web access forms. Ok, enough
rambling, time to plan the build schedule for this week and hopefully the bumps have been flattened out for awhile.    
9/18 The charger build is moving along. This week was focused on organizing the documentation for the charger
wiring, finishing Charger cases for I/O connectors and wiring up the first charger for the build of 5. Getting setup on
the Lab PC was rather time consuming, slowed by an old dos program in use today (too much $$$$ to upgrade
another option for E-Meter users to choose from. I've updated the website to reflect this new item. Time for some rest!
in the lab, so now I've got more room to work with. More bench space allows for more projects to be in various states
background) on increasing the charging power level with a larger isolation section since I have a larger enclosure.
Likely that will be awhile from a product improvement with all the charger builds going on. Time for some rest!
10/1 With boards nearing done, things have been eventful over the last few days. No failures to highlight, probably
more frustration dealing with lead. Seems my dyno pack has seen better days, at year 11 on a 10 year life rating for
the batteries. I bought them used around year 4-5 when I got them ($~30 each from a UPS installation) I'm now having
a few weak ones showing up done at higher charge rates. Seconds after the charger is plugged in, some OVPs are
blinking, throttling the charger back. In order to test out the chargers, I need to dump some power over 5-6 hours of
test time. One thing I can try is to do something similar the Dyno does. Take the DC output of the charger, isolate it,
convert it into synchronous AC and push it back on the line. The drawback there is the transition to the charger
throttling is lost, which is a point of interest. So while that may still work, I have to consider it as an option, or swap out
the entire pack for some 155Ah ones waiting for something to do. Decision time, but I'll have to look into it tomorrow.
10/3/09 After looking at my current pack options, swapping out a few weak ones in hopes of getting a stronger pack, I
decided to swap out the entire pack for a set of Marathon 155Ah batteries. I forgot just how heavy they were. All in all,
I think I shuffled around 1200 lbs of lead today, sure don't want to make that a habit. They're being fed with the
charger at the moment, then its discharge time to see what my initial Ah's are. I had bought these from the same guy I
had converted a golf cart into a mower. These too were in a UPS installation, (think I paid $90 ea in '05) and had only
2 yrs of standby time than the previous pack. Once I establish the Ah,  then I can start to do some charger cycling
tests (looking for increasing Ah's) and finish out the remaining build. I'm tired, its time for some rest...
10/10 Returned today from a busy trip, its sure good to be back. Bringing some parts with me meant I was able to
keep the charger build moving along as well. Good progress all things considered, now this week's focus will be
getting the BatMon / LionMon boards finalized and the chargers wired up. Once that's done I can begin testing, and
with less wasted power (heat). Usually electric utility bills have been in the higher cost rates  (as high as rate tier 6 on
occasion) so having the ability to discharge the 155Ah pack back into usable AC power instead of heat should save a
bunch. I got to watch the meter run backwards with some TV's on plus a few computers, so while this may seem trivial
to some,  it generated a bit of excitement on my behalf. It was short lived though when larger appliances kicked on,
once again putting me back in electric debt. I need some anxious hamsters, a bucket with a motor shaft attached to
the bottom, hmmm, free generator power! If I can feed them for free, is that over-unity? Yep, time for some rest...  
current setpoints over the 120VAC-240VAC range, so I decided to eliminate the circuitry, getting rid of 15-30 minutes
of calibration time on each board. This change only affects the single PFC 1.2kW/2.4kW chargers (not the earlier
1.2kW-only chargers), making them easier to build. Around Wednesday I decided it was time to get sick as a dog and
guess what? It worked... I'm now semi-recovered and even more cautious about shopping with all these sick folk out
there. Hey -little ones- don't wipe your nose with the shopping cart handle! (Haven't seen any adults doing that -yet).
As you can tell my comedic outbursts fly out while the fever is in the taper down stage. Ok, so while things are getting
back up to speed, I've got to pace myself to keep from going back into deep chill mode (103F or more). More to do...  
10/24 Still no word from the PFC IC manufacturer, hopefully I'll hear something back regarding my date code
question...Sent out another dual charger yesterday, with a few more chargers remaining in the pipeline. I somehow
managed to not fully recover so now I'm starting to feel worse than I did a week ago. I'm thinking its coinciding with a
recent shopping trip, but no way to be 100% sure. I'm finding that additional deadtime in my PWM IC's for the charger
is costing me less voltage-wise. The output used to be stiff from no load to full load, now I'm having to increase the
output voltage to compensate for it. I hate it when parts from different manufacturers (but same part number) don't
hate to have to screen parts for parameters to be sure they work correctly... Anyway, there's always more to do...  
11/2 Many things going on concurrently these days, but alas the build inches forward. Then its on hold? What's up?
intrusion or what. But as sure as they are installed and run, pop goes the cap. It can sure make you jumpy once
you've heard them go pop. These caps are designed to be fault tolerant since they heal once the fault has cleared,
the downside is a loss of capacitance (designed for mil spec applications). Ceramics tend to fail short and stay short.
This has been the first time I have witnessed this many caps faulting in any build cycle. While I'm suspecting moisture, I
call to ITW to see what they recommend. Might be as simple as a high temp bake cycle to drive out the moisture. Also
found my 2x sized PFC diode is much noisier, limiting my current output (only on the single chargers). So some noise
reduction is on the way to cure that. Seems to be something lately thats "interesting" and seemingly slowing things
down (that's the part I hate). I'd rather be doing assembly and test than troubleshooting to find the bugger causing the
problem. What can be said is, now PUSHING forward...
11/7 Back to build progress mode. Found my circuit was to blame for popping caps. Duty cycle variation was placing
more voltage than acceptable on the caps. Now with better grounding, things are nice and quiet, running as expected.
Building LionMons takes 4x more time due to the volume of boards to stuff and test. Seems the 4.25V limit has been
abandoned by all EV'ers now so likely its time to change the part spec on that board. Will be testing PFC diodes this
weekend to see if there's a quieter one that will work and handle 2x more amps. SiC diodes are nice, but pricey for a
600V with schottkey like performance. Also not available in a larger package that I need. Back to work!
11/17/09 Since my last update I've been looking over a past circuit change in the output rectifier section of the
charger. With that I decided to make a component change which has become a drawn-out event simply due to the
process of change / document / test / loop to 1 until satisfied and confirmed. Its ended up being an improvement on
the charger isolation section that made things work better with the PFC. A side benefit is it now runs in partial ZVS
mode under full load conditions. Thus far the charger hardware has been good, with changes in the past mostly due
to part date codes and using different manufacturers parts (same P/N but usually requiring minor value tweaks) has
slowed things down, this change is significant to implement on new chargers being produced today. The overall effect
on things outside the box is probably not noticeable compared to older models. So with a minor delay getting a part
sorted out to handle this task, its time to stop chatting about it and get back to work!
11/27 Testing all this week led to deciding on a new core that can handle the circuit configuration. While some cores I
had in-house worked initially, once the charger was let loose to charge a battery pack (2.4kW in a master-slave
configuration) the cores got warmer than expected over time. So with the new cores on order today, they should be
here early next week. Back to work!  
12/5/09 Finally after finishing up updating a charger, its now on the way to a customer. What a relief! So with the
change implemented, the waveforms look good, core heating is low, and I'm working on the remaining chargers (which
are already close to done) and getting the new parts installed this weekend. Back to work!  
12/15 The next charger tests didn't fare as well, primarily due to diode recovery noise affecting the PFC max current
level. It seems that the diode parameter variation is wider (probably since there is no max spec, only typical) so the
mfgr has no maximum limit to call parts 'bad'. By not specifying a max spec, most parts can ship, aside from those that
don't meet voltage limits. For boost applications, there are a few parameters that matter most, but max values are
often avoided on the spec, hence the push for new technologies as silicon carbide. But there, the cost goes up 5x for
the benefits. So the designer typically has to add more filtering just to accommodate part variations.  Back to work!  
12/21 With another 800A system packed up and ready to ship out, the balance of outstanding charger hardware is
nearing completion. Some SiC diodes will be on order to see just how well they will do in the PFC section. Since the
focus will be on getting more 240VAC power out of the charger, the quieter things are to begin with, the better. The
concept is to get 2.4kW/charger section with the dual PFC input configuration. That would equate to 4.8kW on the
240VAC input side! With that much power, internal cooling will have to be upgraded to handle it, so more investigation
will be needed there as well. Opening up the case to be aircooled invites all sorts of stuff coming inside. Back to work!