INSTRUCTIONS:

To find K40 conversion information use the "INDEX", "SEARCH" or "LABELS" sections in the sidebar.

Thursday, June 29, 2017

K40 High Voltage Transformer Autopsy #2

Background

This is a continuation of http://donsthings.blogspot.com/2017/06/k40-flyback-autopsy.html. In the previous post both myself and +Nate Caine tore down High Voltage Transformers (HVT) as part of our quest for knowledge of the details of the K40 LPS internals.
In this tear down the potting material was removed chemically in hopes that the circuit and its components could be kept in tack. The transformer was also sectioned a means of understanding its design.

Donate:

Please consider donating (button to the right of this post).
Your donations help fund additional research, tools and parts that I will return to the community as information.
For other information on the K40-S build use the  K40-S BUILD INDEX with schematics.

The schematics for the LPS:



Removing the Potting

The potting was mostly removed using paint stripper. Caution: use gloves as this stuff is caustic.
Use gloves and eye protection this stuff is caustic
HVT submerged in glass container
The potting removal took nearly 2 weeks of repeatedly checking and refreshing the solution. The potting will come off in flakes which I washed off with each refresh. Unfortunately the stripper de-laminated the capacitors and destroyed their covering so no labels were visible. The diodes had no labels. 
It is probably worth experimenting with other chemicals that might work faster and not destroy the components coating but this worked good enough to get the information we needed.

Discovering the Circuit

The component connections were carefully observed as the potting removal proceeded. This version (and it seems there are more than one) used 2 HV diodes and 2 HV capacitors in a Voltage Doubler configuration. In this case (unlike the previous autopsy) there were no parallel diodes.

Circuit removed from the potting. Overlay showing connection of secondary.

Image result for voltage doubler
Example circuit. These components are not what is used in a K40 LPS

HVT Cross Section and Analysis

After the potting removal step the transformer was sectioned with an abrasive metal blade on a Dremel and then polished on a marble flat plate with 600 grit wet paper until the wire cross sections was visible.

Left: five section secondary. Right one section primary winding's


Primary Winding

The primary winding consists of 40 turns of 21 wire bundles.
Primary winding


Secondary Winding

The secondary winding consists of 5 sections of  +400 turns wired in series.

Primary winding connection leads


One section of secondary winding
The secondary does not have a center tap and is directly connected to the junction of the diodes and the minus leg of the lower capacitor. See voltage double'r example.

HVT Design

The HVT consists of a primary winding and a 5 section set of secondary winding's.

HVT Primary

The HVT primary has 40 turns in which each turn is a twisted set of 21 wires. The primary effective wire size indicates that the primary is designed to handle much more current than the secondary.

HVT Secondary

The secondary has 5 winding sections insulated from each other but connected in series. 

The turns wound on three different secondary sections were counted, one on the first  HVT and 2 from the second HVT. The first count = 499 and the two sections counted on the second HVT = 452 and 471 respectively. 
I don't think the difference in the counts are caused by counting errors. Its seems that the turns on each section are not the same (Rt). I did not further investigate this variance in turn counts as I do not think it will materially change the outcome of the analysis.

Turns Ratio

Using an average of the last HVT 2 sections turns count lets assume:  

  • Average turns per secondary section = 461.2
  • Number of secondary segments = 5
Total # of turns = 461.2 * 5 = 2306 total turns on the secondary

Therefore the HVT is a 40t to 2306t HVT transformer, a ratio of 1:57.65

Estimating Output Voltage

The voltage at the output can be expressed as:
Hv = Pv * Rt * Mv
where:
  • Hv is the output voltage
  • Pv is the voltage on the primary
  • Rt is the ratio of primary to secondary
  • Mv is the voltage multiplication factor
Therefore for every 100V on the primary the output = :
Hv = Pv * Rt * Mv
11,300 = 100 * 57.65 * 2

Calculating HVT Primary Voltage

The typical K40 PS output voltage is specified at 23,000V @20ma.


If we solve the above equation for Pv:
Pv = Hv/(Rt*Mv)

And use it to estimate the primary voltage at spec: 
Pv = 23000/(57.65*2)
Pv = 200 Volts

By inspection of the LPS schematic I estimate the HV buss to run at about 240-300 volts, 40 volts larger than the estimate above..

This error of 40 volts on the primary equates to about 2,306 volts on the output or 10% of the specified output. This error can easily be the result of an error in estimating the total turns across the 5 secondary sections [perhaps each secondary section has different turns] or simply differences in any given manufactures specified HV output.

Primary Current Estimates:

The specified max current output for a typical K40 supply is 20ma. In a transformer the voltage on the output is increased by the turns ratio. So to the current in the primary is larger than the current in the secondary by that same ratio (Rt). 

Therefore: 
Pi = Si * Rt
where;
Pi = the primary current
Si = the secondary current
Rt = the turns ratio

Solving for the primary current using K40 LPS specs and given the output current = .02 amps
Pi = .02 * 57.65 = 1.1 amps

Learning's

A K40 HVT contains a high current primary and a multi-section secondary. A high turns ratio secondary in combination with a voltage double'r  creates approximately 11,300 volts per 100 volts of primary voltage.
This autopsy provides a model of the HVT that more completely characterizes a key component of a K40 LPS... its HVT.

If the above analysis holds true then the following has been verified:

  • K40 LPS are easily capable of voltages in excess of 23,000 volts
  • A K40 HVT's include a voltage doubl-er in its output stage
  • A K40 HVT cannot be tested using a standard DVM because it cannot forward bias the internal HV diodes. Each HV diode is actually a serially connected stack of 20 or more diodes. Voltages that exceed 120 volts might be necessary to forward bias both the diodes in this double'r configuration.
  • K40 HVT's are not repairable

Suspicions of K40 LPS Failure Modes

I suspect that LPS failures fall into these categories:
  • AC plug swapped with the DC plug damaging the supply's enabling circuits
  • The low voltage PWM controllers output shorting, blowing itself and the bridge rectifier.
  • The Bridge rectifier failing under load. 
  • Arc's causing excessive secondary current, opening the HVT's diodes.

What's Next To Do On the K40 LPS Quest

  1. Map the actual voltages in the LPS including the primary's HV buss to further verify the above model.
  2. Scope and capture dynamic views of the internal circuitry's operation.
  3. Scope and capture dynamic views of the Lasers current and voltage while marking.
  4. Review the component specifications and verify that specifications are not being exceeded in actual operation.
  5. Noodle a safe HVT DIY HVT tester. 
  6. Noodle a safe and DIY HV tester

Enjoy and comment
Don


Saturday, June 3, 2017

K40-HVT Autopsy

Background

A more recent post on this subject.

The High Voltage  Transformer (HVT) is the business end of the K40 LPS. To date we have little knowledge of what is inside that "potted" assembly.

That is about to change as the result of the communities contributions to the HV lab and the dead HVT that  +Phillip Conroy sent to me all the way from Australia.

These two samples were first tested using test methods that I devise while trying to find a way that we could test LPS's safely and at lower voltages, with commonly available equipment.
After the tests, one of the samples was torn down with hopes of discovering what was inside the potting.
The one on the right had an autopsy. 

References to Related Posts


Donate:

Please consider donating (button to the right of this post).
Your donations help fund additional research, tools and parts that I will return to the community as information.
For other information on the K40-S build use the  K40-S BUILD INDEX with schematics

Testing

The HVT's that +Phillip Conroy sent were tested using the simple 9V battery test and a ZVS (Zero Voltage Switching) flyback driver design that I have been tinkering with.

Battery Test

I tested both samples with the 9V Battery Test and both samples passed.

ZVS Flyback Driver

I tested both samples with the flyback tester made from ZVS driver.
One passed and one failed.

The ZDS driver uses a re purposed  SainSmart 5V~12V Zero Voltage Switching ZVS Induction Heating Power Supply Module + Coil Power Supply heating power supply module as a HV driver. The K40 fly-back's primary was connected to this drivers output and a spark gap was used for visual verification.
ZDS driver connected to a K40 HVT. Spark gap created between HV lead and grnd.


Testing Conclusions

  • As I suspected the battery test will not identify all defective HVT. Certainly, if it fails this test it is bad, however if a transformer passes this test it still may be bad..
  • One of these samples is clearly bad (no output using the ZVS driver).
  • The other is either good or the ZVS drive also cannot find all types of failures. It could be the case that the ZDS driver does not develop a high enough (voltage) drive on the primary and therefore does not stress the outputs HV section.

Fly-back Autopsy

With a known bad sample I could now proceed to try and disassemble the potted HVT module.
After removing the outer shell with a dremel saw and chisel I tried various methods:
  • Boiling (NOT)
  • Heating with heat gun (NOT)
  • Real "hacking", chipped with chisel but pieces broke off like shards taking the circuit with it
  • Dremel tool, worked but would take many days and lost of bits.
HV diodes and what look like caps exposed and sheared off with the chips.
Some Dremel work to try and expose the circuit


Toaster Oven

I heated the unit to 450F and the material would flake off, rather than chip when stabbed with a pointed. While it is hot the potting becomes a hard chalky-like consistency. With careful poking I got parts to release. Unfortunately by the time I discovered this technique some of the parts had already broken and circuitry came off  with the chips making it harder to guess the circuit. At least I have one method that works and I now know where the parts are.

Results of Tear Down #1

I learned the following from this tear down.
  1. The HVT contains more than just primary and secondary winding's.
  2. There are 2 additional component types inside the HVT and the circuit possibilities are.
Notes: 
  • The additional components are across the HV (RED) and ground (BLACK) side of the secondary.
  • The black components measured open. HV diodes measure open when tested with a DVM. 
  • Split Core (left-right halves)
  • Secondary coil (bigger)
  • Primary coil (smaller)
  • HV capacitor (blue), 2x
  • Diode assy (black), 3 pairs
  • Stud (copper). 2X

A Much Better Autopsy & Analysis

Since my attempt to spill the guts of a K40 HVT +Nate Caine expertly cross sectioned a HVT. One of two from a 80 watt machine. By inspection the single HVT from the pair looks to be the same construction as my 40W sample, except for the number of diodes in parallel.

+Nate Cain's Autopsy disclosed that the circuit that I could not identify is a HV full wave bridge made up of  4 sets of 3 diodes in parallel. Each diode is made up of 20 diodes likely to attain the roughly 20,000V the HVT requires.

One difference between the K40 and K80 transformers is the # of diodes in parallel. The K40 has 2x in parallel whereas the K80 has three.

Next Steps

Explore a chemical alternative to removing the potting so that we could potentially test the internal components of bad HVT's.
Cross section the other K40 HVT sample like +Nate Caine did.
Cross section one of the diode pairs.
I think we certainly have enough information to understand the HVT module, the last piece of the LPS puzzle. Now we need to find out why these fail so often.
Update the schematics to coincide with recent community findings.


Enjoy and comment
Maker Don