RDWorks Learning Lab 250 Deep Dive into HV Power supplies
HTML-код
- Опубликовано: 16 апр 2023
- A few sessions ago I had a great idea to transfer the almost instant response times of the RF CO2 laser to the glass tube laser. The main response delay with a glass tube is whenever you switch the tube off there is a restart delay. ALL manufactureres promise a switch on response time of less than 1ms.There is no physics reason for the pixel to pixel current flow to be other than almost instant, provided the glass tube never switches off. If you switch the beam off then there are TWO delay issues. First is that that 1ms start up to create the high voltage and secondly the tube has a weird performance characteristic called preionization which adds more delay. With dithered photo engraving the tube is continually switcing on and off because black pixels = ON and white pixels =OFF. White(greyscale 255) will ALWAYS act as a OFF signal. Thus, after and OFF, subsequent pixel demand cannot be satisfied until the tube is back to full power. My idea was to dither an image (just black and white pixels) and then convert it to grayscale where black pixels =0 and white pixels = 255. I then adjusted the image output so that I removed the 255 white pixels by limiting the image to 254. No one except the controller would notice the missing 255. This in fact, does prevented the tube from switching off during a scan and shoiuld have given me the instant current flow changes I was expecting.
BIG disappointment. I turns out on closer inspection of the signals, the controller is very happy to react instantly and never issue a "switch-off" signall but the current control circuit within the HV power supply is VERY heavily damped and what comes out is but a shadow of the demand signal going in.
I discussed this issue with Cloudray to see if they were aware of this performance problem. They were not. My next question asked if there were any differences between HV power supply manufacturers.? Again the answer was unkown. However it prompted Cloudray to offer me two different HV power supply designs to compare against my basic MYJG design. The results were rather suprising 
Наука
Thank you! Very interesting! It was a brilliant analysis of the behaviour.
Very interesting. I think the massive peak at the start is because the tube needs a high voltage to strike but once the gas is ionised it runs at a lower voltage. So if there is any capacitance in the circuit you get a big current spike. If the current is too low the discharge stops and the voltage has to build up again to make it strike again. That gives the unstable "pre-ionisation" stage.
The slow response to PWM changes is probably because the PWM is converted to a control voltage with an integrator, most likely a simple RC filter. Perhaps a better system would use a DAC instead of PWM, or a much higher PWM frequency and a faster filter.
Many thanks
That is a nice simple explanation of the behaviour. We cannot easily change it but understanding that different power supplies manage this unstable phase in different ways has allowed me to dramatically improve my photo engraving speeds to almost match the 800mm/s capability I can achieve with my RF machine capability. However, the sluggish pixel to pixel respponse will always produce a "smoothed" 3D engraving result and a fuzzy grayscale engraving.
definitely waited for this topic! :) Trying to see if there are ways to improve response speed for my tube too
Thanks for this review. I just installed a HY-Z Z-80 and am quite impressed with its performance. I was curious to see that the voltage is not constant, across the range of percentages. This one goes, from a low of 13.4kV, to a high of 25.7kV, before settling down to above 35%. The response time is significantly higher than the stock MYJG supply that came with laser.
Hi James
Your curiosity is justified but dont worry the voltage is never going to be stable. The purpose of the high voltage is to break down the nitrogen gas which represents about 10% of the gas mix in the tube. All gasses are naturally non conductive ie they have a VERY high resistance,. When you excite the nitrogen molecoles with VERY high voltage it causes some of the electrons to break free and it is those free electrons that carry current. Simple Ohms law says Voltage = Current flow X Resistance. So when the gas "resistance" is almost infinite there will be no current flow so the voltage goes higher and higher until it reaches a point where th nitrogen starts to ionize (the pink glow in the tube) and current starts to flow because the resistance of the gas drops to almost zero.. This is bad news because low resistance and high voltge = massive current flow = exploding tube!!!! That is where the other function of the power supply takes over and limits the current flow to a controllable few milliamps. So back to Ohms law and now we have low resistance ionized gas (lightning in a bottle) and low current flow so it only requires a low voltage to maintain the ionized gas state. The trigger voltage for the tube depends mainly on the length of the gas column. Thus higher wattage tubes are longer and require more voltage to create ionization. However once ionization is established the high voltage will drop from say 25kV to maybe 15 to 17 kV.. The only important thing to worry about is the current flow through the tube and that is something YOU have control of. %power is not watts but current flow through the tube. Look at your power supply spec and the HY-Z Z80 can dliver up to 28mA. That means that 0% power = 0 mA , 50% power = 14mA and 100% power=28mA. If you have an 80 watt tube the manufacturere will specify a working current, probably 26mA max. That means you are allowed to use up to 26/28= 92% power in your programs.. The watts outputf rom your tube will never be linear so an 80 watt tube will not deliver 40 watts at 50% power, it will probably be closer to 60 or 70watts. The HYz series of power supplies appear to have no different response times to any others EXCEPT in the way it manages the preionization zone. For most normal applications you will see no difference between any HV power supplies . but my careful examination of the preionations zone allowed me to exploit a property that other power supplies do no have. Its brilliant for photo replication of dithered images.
@@SarbarMultimedia when I say response time, I am referring to the power supplies ability to rapidly turn on and turn off, especially in the Pre-ionization zone. Due to, what I consider improved response, I am able to photo engrave at speeds of 300mms, where, with the stock MYJG power supply, anything above 100mms, and I would start loosing resolution in the picture engrave.
For cardboard/paper cutting - do you have a min power set (that is lower than the max power)? The reason I am asking is because if there is a min power, then it will be lower than the pre-ionization power. For me, my accel values are set low in order to get straight cuts, but that in turn requires me to have the min power lower to not burn the corners. What do you think would be the best solution in this case?
Hi
Max and Min are equal because this is not normal constant power cutting. Enen though you still slow down forthe corners, the very short high power pulses do not cause burning because they are always instantly vapourising the card.
Reducing acelleration to achieve staright cuts sounds like a mechanical issue. If you wish to contact me privately so that we can try to sort this with email and pictures/video then use the following form
forms.zohopublic.eu/ndeavorlimited/form/K40XtreeemLaserCutterContactRussSadler/formperma/k2Cn0QN5ChpazfTMAUw25lZ-FKpjZa96TQWHjv3ntOg
Hi Russ, thankyou for these very interesting videos! A question please, the maximum current for a given tube can be limited by adjusting a trim pot on some HV PSU's, is there any disadvantage in ignoring that trim pot and instead limiting the current by reducing the allowable PWM duty cycle max in DSP controller settings (Laser max power% in Ruida)? Or is it effectively the same thing? Does the trim pot simply internally reduce the voltage of the PWM signal after its been converted back to analogue inside the PSU? My head tells me that the option of trimming the PSU and utilising the full 1-100% PWM pulse width/duty cycle range to relay the power control signal would result in better resolution when engraving varying tonal burn intensities in images? Penny for your thoughts.
Hi Nic
A glass tube makes no use of the PWM because it is a CONSTANT current system that never switches on and off during running. It is only RF and diode lasers that can exploit the the pwm control. I believe it is preset to 20kHz to control the stepper motors. Yes you can use the PWM output from the controller to controlthe %power into a tube but it can be confusing. Ruida recommend you use the analog output at AN1 at CN5 for glass tubes. and the PWM output for diode and RF lasers. However, if you use the PWM output for a glass tube it just converts the analog output into a digital signal to encode the %power. As soon as that digital signal arrives at the HV power supply andinput filter convrts it back to analog. So ther eis zero control of your glass tube with PWM..
The next issue you have to come to terms with is understanding what % power means. It's nothing to do with watts. It all about the max output current that your HVpower supply controlls. 100% power = max HVpsu output and it is linear all the way to o%power=0mA. I So if you have a power supply that delivers 30mA ma and your tube requires 24mA max then your max current should not exceed 24/30=80%. see no value in fiddling with the PSU output to match that allwed through your tube. If you have an ammeter fitted its just a matter of running it at various powers to find the % power that allows the max allowable current through your tube. Lets say ist 80 % - 24 mA then you can set 80%max in the vendoer settings and never fear overdriving your tube. Ah, I here people say, but that means I lose some of my power control because I am only ever using 80% of the range. Hmm .... is that REALLY a problem? %power can be set in increment of 0.1%. That mens you have 800 power increments to platy with Yes you may need close control of low power if you do photo engraving but in 8 years with 3 different tube powers I have never use a 0.1% power increment. I have never experimented with that control so I do not know if it is a proportional gain control or just a peak current limiter.....which is exactly what the Max %power does in the vendor settings.
If you look at any of my scope traces of the current flow , the only time it goes bonkers is during preionization and that is totallty nothing to do with PWM. The PWM I use in these tests is so that I can see the 50 microsecond 20kHz digital PWM signal that the controller sends out. I can also instantly see the % power (current flow) demand that the controller is asking for (the PWM duty cycle) and the difference clearly shows that all these HV power supplies are pathetically slow. The only reason I am able to get rapid response for photo replication is by staying in tha bonker out of control preionization region where the performance is accidently close to that of a PWM controlled RF laser.
Best wishes.......Russ
@@SarbarMultimedia Thankyou Russ for your prompt and detailed reply, very informative. You have left me with a lot to think about. If I get the chance I will try to run my own tests to find the preionization zone on my system and test for more responsive engraving - fascinating stuff.
Thats what I thought that all those designs are 100+ years old haha. So, there is still no suitable good power supply for regular etching/engraving, right? As far as I understand the last one you tested works the best, but only in very low power modes, right? I have to etch my materials at at least 600mm/s and preferably at 1000, and this delay issue is killing me...
Hi Vlad
I typed a reply to your comment but it seems to have been posted to a paralleml universe. Did you ever see it?
@@SarbarMultimedia haha, you guessed right - not at all. I do see this one only. 🤷♂
Hi Russ, have you found any difference in the laser output, response time or pre-ionisation when connecting the power supply to the the PWM and Analog signal from the controller?
Great question, but No. That's because the PWM coding is a staright A to AD conversion intended by Ruida for RF machines. However there is no delay caused by this encoding and it works perfectly for glass tubes. as well as RF tubes. When this digital signal reaches the IN terminal of the HV power supply it is immediately converted back to analog by passing it through a an RC filter.. That first high current preionization spike is being "arrested" (eventually )by a feedback circuit within the power supply and that is why we see the typical capacitor charging decay in the first and subsequent pulses, Note that those pulses are not in synch with the 20 kHz PWM signal.
Best wishes
@@SarbarMultimedia As you know i have watched a lot of your videos and like to recreate some (not all) of your experiments, some are successful others not so much, but one thing I have not be able to get is the pre-ionisation phase of my tube. my tube is a grade A tube 120W Reci which lights up at a minimum setting of 12% power, there are only a limited number of contributing factors but my machine is built with the connection from the Analog signal and I'm just wondering if I swap this the PWM signal will I have more luck?
@@precisionlasersolutions7996
Hi
I an not a fan of Reci tubes , despite their revered reputation. Nice engineering but some weird performance characteristics. They have only ever produced high power tubes ( 90 watts and above) that can be considered as CUTTING tubes. I have direct experience of 8 different Reci tubes from W1 to W8 and all but 1 have displayed a strange slow beam start for scanning (over maybe the first 2 or 3mm of scan) where the beam gradually builds up power. You need to be running at maybe 40% or more before this characteristic disappears and you get a uniform burn scan line. This annoying characterisic may be a feature of the special Reci HV power supply because the singe machine with a basic MYJG power supply where the Reci tube was an upgrade, din not exhibit this quirk. That aside, ALL glass laser tubes will have a preionization zone regardless of quality, power or manufacturer. in fact here is a quote from the Reci operating manual that sound like gobbldygook
"Taking engraving as an example, 4mA is recommended to produce pulse with high frequency under preionization state for better engraving."
The best way to find the limits will be to cut some thin card and LISTEN for it
Here is an example of the amazing benefit of cutting card with the tube in preionization. Listen to how "hissy" the cut is. If you can't see it then do some card trials at different low powers and try to find the upper power limit by the sound of the cut.
ruclips.net/video/S_2NOCpACVc/видео.html
Best wishes
Russ
When i replaced the shorted PS ( internal arc).
I needed one quick so i bought a blue unit.
Installed it loose as the internals of the machine needed to be reconfigured for it to fit. So i ordered the correct black unit.
When that came in i swapped for the black unit. My wife and me agreed the blue just seemed to have less problems.
Only difference in spec i could find was the kV was higher on the blue unit.
Cloudray does not make the PS. They just rebrand them.
I am sure the power supplies also have similar quality issues to the chillers. Some are fakes.
Digital kV meter.. Garbage.
Hi
The kV meters stopped working when I put my restsistance in line to measure current flow but the analoge ammeter didn't. That says a lot about the sensitivity of it's digital .
operating principle. It was only this minute detail view of the internal workings that allowed me to find performance I have never seen before.
There is no such thing as a quality HV psu. Thery are all made to the same design principles with the same key flaky components. The high voltage flyback transformers are the same in all designs.. They factory soak the units for a period to weed out the instant failures but there is no guarantee of longivity by paying more. I have PSUs that are 7 years old and a couple that failed after 18 months and 2 years. It does not matter what the kV output is, it depends on thr length of your tiube mainly. There needs to be sufficient kV to ionize the nitrogen in the gas mix , after that happens (at say 22kV) then it instantly drops back to a lower value (say 17kV) for normal running. If the tube fires then there is enough kV, it's as simple as that.
Cloudray choose there suppliers carefully and enter into official rebranding deals with them. Reliability and quality are significant factors for Cloudray and if their returns/field falure reports exceed very low percentages, that product line is withdrawn.. The Z80 , specifically sold for EFR tubes (why ? I don't know) was by far an away the best overall performer. The digital meter was pathetic also.
@@SarbarMultimedia gunny thing is.. Thr blue and black unit come from the same manufacture.
So the Kv is only relative to the initial firing of the laser? I couldnt find any information explaining specs of the power supplies.
Trotec and Boss do not use the same PS that low budget or Cloudray machines use.
The digital meter is a novelty. Not actually useful. I quickly removed it and put the analog back in.
@@SarbarMultimedia The PSs with remote ammeter connection have a 100ohms resistor in series with the return line. With your method of measuring the current, you short cut this resistor. The resistor input is exposed on the RJ45 socket (3rd pin from the left, 6th pin is GND). You can use those to measure the current.
@@GlimpsOfScience
Mant thanks, I greatly appreciate that information. I did realize that my resistor was disabling the digital ammeter but had no effect on an analoge meter. Before I put my resistor in line, I asked the technicians at Cloudray where I could measure current flow. Sadly their response was rather tardy and so I made my own decision thanks to Ohms law.. In principle the 100 ohms will make no difference to my readings or any current feedback circuits within the HV power supply because the current flow will be the same through all resistors connected in series. The Z80 PSU was connected to its display with 4 wires and obviously had it's own sensing resistor because that could not register mA either. I didn't see this as an issue because I could see the current flow through my resistor as a voltage on my scope and knew I was not overdriving the tube. However, it is rather alarming to see the volume and magnitude of NATURAL overdriving taking place that would never be seen on an analog or digital ammeter. So much for us being cautious about overdriving our tubes!!! Note. There are no such issues with RF tubes because they works with a fixed current flow and there is no preionzation to worry about. That is probably the main reason why RF tubes have a longer life.
AFTERTHOUGHT: My knowledge of electrocs/electronics iis very basic but it sems illogical for the digital ammeter to be disabled. Current flow through the 100 ohm digital sensor has remained unchanged so why does my preceding resistor have an effect?
@@SarbarMultimedia Your way of measuring is by no means problematic and absolutely correct. The 100ohms resistor is located between your resistor and GND. The GND of your oszilloscope connects to the input of the 100ohms resistor. With a standard oszilloscope, GND is the same as the GND of the PS. Now both sides of the 100ohms resistor are connected to GND and the voltage over this resistor is 0. I don't know about the Picoscope, though but your experience seems to indicate that it is the same.
Hello Russ I would be interested in one of your Dohickys so per you directions in an earlier post on the Dohicky I can be contacted at steverichab next add an at symbol then comcast followed by a dot and the net. Please send directions on how I may be able to acquire a dohicky.