PrimeWeld Tig225 Cooling Fan Mod

[keeena]

If you have the PW Tig225, then you know the cooling fans are ridiculously loud. They are also on all the time (vs. on-demand) which is the real kicker. Decided to do something about it.

The PW has 2 fans which pull air in from the rear. 24v, 120mm x 38mm deep. Both fans pull air in from the back and thru a tunnel lined with heat-sink fins. I couldn't find anything on the web for the exact model fan; this was the closest match I could find: link. Given the CFM and wattage of the fan I did find specs for, I'm guessing that the PW fans are at least 120-140cfm. This means you do not want your usual or cheaper-brand "quiet" PC fans - these will not push enough air.


My phone's sound meter shows about 72dB 2 feet from the rear, 66dB 2 feet from the front.


The fan I settled on is the Noctua NF-F12 iPPC-24V-3000 (link). It is 120cfm and designed for pressure applications (radiators and heatsinks). They are rated at 42db (fairly loud compared to the usual "quiet" fans) and can be speed controlled via PWM. For now I decided to run w/o PWM.

The swap is pretty straight-forward. Besides the fans, I bough new plugs to avoid cutting the original fan wiring (Plug and Pins). I couldn't find a male plug to match the Noctua's 4-pin connector (they do exist, just couldn't find a a vendor), so I cobbled together a solution using standard PCB header pins for now.


I also designed & 3D printed a new fan duct. The Noctua fans are 25mm deep, so a deeper air duct is needed to keep the airflow directed thru the heat-sink tunnel. Model can be found here: link.


Install Tips
Remove all screws around the top and sides of the sheet metal cover. Remove the 4 feet but do not remove any other bolts on the bottom of the case. There is a ground strap under the top of the case which also has to be removed. Then the cover will come off. You should also remove 4 screws (two on the top, two on the side) which attach the back plate to the frame. This is needed to get a little clearance so you can sneak the fans out. The hardest part is removing the silastic on the fan plugs. You really want to take your time to scrape off what you can and slowly/gently ply the plug off the header. The plug's lever/tab will also likely be glued to the header - some gentle prying and scraping will get it loosened up.

Final Result
~60dB rear, 56dB front. Although still noisy, it's a significant drop.



That said: I'm not done yet.
I'm going to make a duct for the top fan. The original fan rested right against the top heat sink tunnel so no duct was needed. The new fan is 13mm shallower, so I may print a 13mm thick version of the bottom duct.

I also plan on adding a thermostatic control, but first I need to do some testing to get an idea of internal temps so I can figure out thresholds. A more thorough approach would be also spin up when torch is on. This would require a custom solution (ie. arduino or similar). But spinning down to 25-50% will make this unit pretty quiet. I'll resurrect the thread if/when I get to it.

And a bonus
I was never a fan of the cheap ground clamp or length (10'). When I went to upgrade some bits, I realized that the ground was worse than I expected. The wire measures as 4gauge which is technically too small for a DINSE 35-50 (and pushing it for the max amperage of the PW). I was able to pull the wire out of the DINSE connector without much effort. The DINSE set-screw was turned in so far that it was no longer engaged with the threads. Yikes.

I bought a quality 25', 2gauge wire (link) and a Tweco 300a clamp. Made some heavier copper bushings for the DINSE and clamp out of copper pipe. Much much better. The Tweco takes some serious muscle to open...should have Popeye forearms by spring.

 

[cathead]

I liked your post and your quest to quiet down the machine. Just remember that the enemy of the IGBT and solid state devices in general
is heat. I'm pretty sure that would be why they don't incorporate thermostatically controlled fans. Using a thermostat will have a negative
effect on the duty cycle. It's still a free country(sort of) so you can do as you please of course. It's something to think about though
when making modifications. Better ground wire and clamp certainly is an improvement. I was at the local farm store and see that
the price of copper welding wire has doubled here in only a year.

 

[graham-xrf]

Great write-up! Comprehensive work, measure know, fix.

There is some hardball stuff to know, and apologies here if it's what you already know. I don't know how calibrated phone apps can be, but here, that does not matter. It's the relative improvement you wanted to measure. The readings you get correspond to noise power changes of 10*Log10(Ratio).

Rear noise improvement 11.3dB. That makes ratio 13.5, meaning the noise power is 13.5 times less.
Front noise improvement is also good, 10.1dB. The ratio is 10.3 times. Basically the racket is about 10% of what it was.
Our hearing does somewhat compensate against this. The improvement will not be perceived as being as much as that. The brain turns up the gain!

About the cable length. The copper resistance increases proportional to the length, and decreases inverse proportional to the square of the copper cross-section area. The resistance of a connection can often compete with even quite long leads, so going to 25' is perfectly OK, provided you use the thicker gauge copper, and take care of the connections. Even a small increase in copper gauge has a dramatic effect on lowering the resistance, because of the value being squared to get area, but the connections quality is usually more important.

25' may have 2.5x the resistance of 10', but that is covered by the thicker copper, and easily covered by fixing the connections, so you did the right thing there. A connection resistance only 0.1 ohms is going to drop 18V when you start running 180A, and put 32W heat into the connection. It will get hot, and start to get worse!

@cathead is right about the temperatures of heatsinks, whether FET or IGBT. Be careful about messing with heatsinks. They often have high circuit voltages on them, and the junction temperatures in the semiconductors are a whole lot higher than the temperature on the heatsink fins.

 

[Firstram]

I bought a quality 25', 2gauge wire (link) and a Tweco 300a clamp. Made some heavier copper bushings for the DINSE and clamp out of copper pipe. Much much better. The Tweco takes some serious muscle to open...should have Popeye forearms by spring.
View attachment 387093 View attachment 387094

I have done this to all 5 welding machines at work, makes life so much easier!

 

[keeena]

@cathead - Totally agree. Changing from always-on to on-demand cooling requires quite a bit of understanding about how quickly the circuit heats up, operating temp threshold, temp recovery, temp maintenance, etc... Its the main reason I didn't bother tackling it. Happy at the moment with the modest improvement.

Should be stated that this type of mod is a YMMV, proceed at your own risk!

@graham-xrf: Yeah, I don't put much stock into the phone's SPL accuracy. You're exactly right that I was just looking to understand relative improvement. It is funny how the dB improvement doesn't seem as good to the ears as it does on paper. Brain gain.

Also agreed on the grounding. At first I was just looking to get the wire gauge more in-line with what it should have been from the factory IMO. But then I was particularly concerned after noticing the less than ideal connection on my DINSE connector. While on the topic: what do you guys think about using electrical grease on the termination points?

 

[graham-xrf]

Grease over the connection, put on after, is a corrosion protection, and know that it is the positive DC terminal that would be prone, because like all corrosion, it is voltage-driven.

Here we do get picky. Grease between the metals is not wanted. Grease is an insulator, and the connection, if it has grease in there, will only connect at little points forced together to displace the grease. Clean the connect metals with alcohol, or something that completely evaporates, and when tightening, then try to rotate relative to each other, to slightly deform metal surfaces into intimate contact or even cold-weld somewhat, then cover with the grease after.

You will have noticed the negetive pole on a battery never corrodes, and the old positive-grounded connection cars were only done that way to make them rot faster as a part of planned obsolescence policy, with various false rationale. Folk have long discovered that covering the terminals in grease helped the positive to not end up as white mush. That is where the original motivation came from.

 

[keeena]

@graham-xrf : I didn't know/think that corrosion would only affect the positive terminal? But even so: what about when in AC mode where the leads are alternating between pos and neg? Agreed as far as grease itself not being conductive and better to be applied after connection is assembled.

 

[graham-xrf]

@graham-xrf : I didn't know/think that corrosion would only affect the positive terminal? But even so: what about when in AC mode where the leads are alternating between pos and neg? Agreed as far as grease itself not being conductive and better to be applied after connection is assembled.

Ever since more years ago than I care to admit to, when I would have to charge all manner of batteries in a place where the acid fumes were thick enough to make one sneeze, I got to know about batteries, DC, and corrosion. AC terminals from 24V 400Hz rotary inverters corrode in pulses as the terminal cycles postive, and halts while it is negative. The corrosion is shared. Both ends corrode equally, but corrode they do. The potentials you get pretty much depend on the electrode metals in there. There needs to be moisture in the air, and then, the terminal will start it's oxidation chemistry.

There is a reason tons of sacrificial zinc anodes are attached to the sides of ships. It is so they can preferentially corrode away, while protecting the ship iron from rotting away so fast. If you look at your car battery, it is usually immediately evident which terminal is benign, and which one is having a hard time. This corrosion is driven by the battery voltage, which is way higher than the potentials in the electrochemical series. Connected to metals in a conducting liquid, like an electroplating bath, you can make one terminal disappear, dissolved away, and have the other become plated over, accepting the material. Corrosion into wet air does not quite work like that, but it is the way even lead terminals become covered in white oxide, and rotted full of pits

Covering with grease excludes the access to oxygen. It's protected from corrosion by air isolation.

 

[keeena]

I tackled the same sort of mod with the PW water cooler yesterday. I only just started using the cooler and yes...it is crazy loud. Quite a bit louder than the stock Tig 225.

Front sound level. The initial volume in the chart is the Tig 225 only; then I switch the W/C on a few seconds in. This was taken at roughly 2' from the front of my welding cart.


Rear; also about 2'. Both the Tig 225 and W/C are on.


There's another post on H-M (here) and ToT's video; pretty similar but figured I'd share a couple specifics to the PW unit and how I decided to tackle.

The PW has a 3-row radiator vs. the 2-row on the HTP Arctic Chill; less clearance between the rad and pump motor. You will have to remove the fan adapter; luckily it just slides off the motor shaft after the fan has been removed (fan screw is LH thread). Even with the fan adapter removed, 1" / 25mm is the thickest fan you'll fit. The shaft is pretty close to the fan blades; maybe about a 1/4" of clearance.

I used Noctua "pressure" fans again, but slightly bigger at 140mm. Used 24v fans. Mounting is straight forward; I used #8 nutserts in the bottom and just zip-ties for the top.


I mounted the power supply off the floor of the unit to help avoid it getting wet if the reservoir leaks or overflows. I used some aluminum brackets riveted to the reservoir bracket. I removed the bracket since I didn't trust myself drilling it in place. It's possible to remove this bracket without disassembling anything else; just tilt the front panel & res a bit forward you can shimmy/tilt/finagle it up and out the top side. Just take your time and avoid snagging hoses & electrics.

You'll need to place the rivets near the very edges of the bracket so they don't interfere with the reservoir. The res has curved edges which gives some clearance for the rivets. The PS brackets were riveted on AFTER the res bracket was back in place but not yet bolted in so I could access the rivets towards the front of the unit. You will not be able to get this assembly in/out with the PS brackets attached (at least not without detaching hoses).



PS installed. I tied the power supply into the 120v output side of the contactor. The PS I used can be found here if interested. The fans are PWM, but you can run them on full voltage and not hook up the sensor or speed input - they just run at 100% in this case. Flow out the back of the unit isn't quite as high, but I think should do the job fine for my use.


I didn't take sound level measurements after, but WOW...SOOOOO much better. I couldn't hear myself think with the W/C on...it was that loud.

 

[keeena]

One more thing: the PW W/C doesn't have any indicator light to let you know its on (the HTP does have a lighted switch). I bought these and I think they will fit but not 100% sure (they are still en-route). I'll update if they are not the right size.