First impressions of New Kemppi Master-M 358G w/Watercooler (Loads of photos)

[Munkul]

I’ve just had a look at the user manual, I’m not sure I understand. The arc length controls and double pulse adjustments are all documented as being adjustable parameters in the menu system. They also have several synergic lines for CuSi in the pulse workpack. It’s everything I would expect to see and adjust.

Is it just a case that they’re missing on your machine?

Could the update have corrupted the memory somehow and ruined your machine? It certainly seems like a possibility.

Have Kemppi UK not been any help?

 

[superclarkey]

you can of corse trim the welder but as its a open loop with static look up tables as soon as your gas is wrong and trying to move too far away from program its not sensing and is just a VC machine.

The issues for me isn't just the machine, its Kemppi themselves assuring the seller it can do CuSi to FE, it can't, either misinformed or do not understand.

For me it was a big investment in a small workshop and 3 years on as its about to come
out of warranty I feel robbed at this point, sure it will be fine for someone likely doing heavy welding but at the same time they are the people that will go for the x range not this..

so I fail to see who this welder is for, and it is like a gimmick, it smells and feels like it should be a trickled down version of top end machine but its not...

this is why you can't add the programs to this like the x series as it lacks arc lengh closed loop sensing, it presumed constant perfect conditions...

yes you can select a program and modify it to a small window... but its not just feed rate and voltage, there is allot more in that program so if its for cu to cu, you can't turn it down enough to work, it will always be way too hot... and without a program no pulse or double pulse can be used...

one thing you do allot in automotive work is being in bad positions and arc lenth isn't constant, I wanted a sensing machine that if when welding around a tube out of position it would help keep constant heat/weld input..

I have borrowed a Fronius machine a few times and realised how wrong I was buying this Kemppi.

Your right to read into how upset I am, I'm loyal to Kemppi and always pushed them, got banners in my workshop and have stickers on ramps and surfaces, so to say I really didn't give them a best shot is an understatement... but the reason they don't want to help me (the conclusion i have come to anyways) is not because they are clueless, far from it, they know what I'm saying is true and its unspoken...

Thing is for most people they will never know, and tbf if it wasn't for the CuSi3 to Fe program not being there I wouldn't have gone down a rabbit hole learning what they do not say in plain english, which to me has caused serious mistrust and the fact they made a gimick welder that was low hanging fruit to catch people like me who doesn't need a welder designed for a ship yard, just wanted a customisable welder that was for a user who was willing to try and get the best out of it.

I was clearly also upset after speaking to Kemppi about the machine being able to be addressed via the widi and as you say being able to remotely control it via my phone (which was amazing, meant i never had to touch the panel and deal with the horrible tactile knobs randomness) they swiftly took me saying how amazing it was an went and released a patch to their firmware to block it... so now I can't use the wifi for anything whatsoever, and can't connect to the machine... felt betrayed and a massive f-u, for me testing the machine out.

Kemppi reps stop communicating with me shortly after I bought it... I have chased up a few times and no reply, the supplier has always chased and same thing, they say it can do it and thats end of the conversation.

I'm not the kinda of guy to sit and argue, I will ask and ask if I get no where I give up.

I now need to decide how I'm going to get out of this machine without loosing **** loads of money, I'm angry at myself as its not only cost me bad, it totally screwed my business plan and what I wanted to do... But Kemppi got paid, so what do they care if they caused a small business owner to be unable to go buy another expensive welder to replace this.

Yes I'm upset, yes I've had 3 beers, I will get over it in a few days and go back to ignoring it in the corner and just Tig welding everything..

they f'up though, I love my mls2300 tig and I would have replaced it with their master tig and likely dropped similar money but they can go f-off now... they have burnt me and I will spend the rest of my life telling people who respect my
opinion what they tried to do and how
they simply didn't care.

Last thing I will say, Rapid Welding have always been very good to me, I don't blame them whatsoever in this, I've spoken to Kemppi direct and got the same brush off, so they have had the same, but also the such small % of people actually buying the 358 with posh screen is so small compared to people buying the basic model who don't care about such things... also it was me that followed the release of the 358 and I waited for it to be released for 6 montjs and I followed every press release and read everything they put out... nothing mentioned anything I later found out, was very much made to look like it had very similar if not the same...

I will get over it, but its hard because I blame myself for trusting them.

 

[superclarkey]

I’ve just had a look at the user manual, I’m not sure I understand. The arc length controls and double pulse adjustments are all documented as being adjustable parameters in the menu system. They also have several synergic lines for CuSi in the pulse workpack. It’s everything I would expect to see and adjust.

Is it just a case that they’re missing on your machine?

Could the update have corrupted the memory somehow and ruined your machine? It certainly seems like a possibility.

Have Kemppi UK not been any help?

sorry missed this in my rand response above as this appeared on page two lol

please excuse this but I rant and don't explain myself well, so just asked chatgpt to compile my findings as I have been going back and forth feeding it documents and information.

"
It’s not a software fault or anything missing on my unit — the Master M 358 just doesn’t give full pulse-waveform control the way the higher-end Kemppi systems do.





Here’s what’s really going on inside the machine.


The M-series uses what Kemppi call open-loop synergic control.


Each program (Fe, AlSi, CuSi, etc.) is built around a fixed waveform set in the firmware — the ratio of peak to base current, pulse width, rise/fall time and background level are all hard-coded into that synergy curve.


The front-panel parameters (arc-length trim, frequency, and “power correction”) don’t rewrite that waveform — they simply apply small scaling offsets to it.

• Arc-length adds or subtracts a few volts from the programmed curve.
• Frequency alters the time base (how many pulses per second).
• Power correction raises or lowers overall energy, moving both the peak and base up or down together.
• Dynamic/inductance just changes the ramp rate between the two levels.

None of these expose the actual duty ratio or let you set independent peak/base current values.


At best you can influence them by about ±10 %, but the underlying current–time relationship stays the same.


That’s why you can turn the knobs and the display shows change, but the weld pool behaviour doesn’t vary as dramatically as you’d expect from a true double-pulse system.





On Kemppi’s FastMig X, X8, and X5 Auto platforms the control architecture is completely different.


Those units have closed-loop arc-length feedback and a programmable pulse engine where the waveform parameters are variables, not constants.


That’s what allows you to tune peak, base, duty and slope directly, and it’s also why they can run adaptive functions like WiseFusion Adaptive or WiseBraze.


The M-series hardware doesn’t have that feedback path or that level of firmware access.





The CuSi lines in the M 358’s pulse work-pack are another example of this limitation.


They exist, but they’re calibrated for CuSi-to-Cu transfer, not CuSi3-to-Fe brazing.


Because the arc energy is too high and there’s no low-current modulation phase, the wire melts the steel surface instead of wetting it — effectively turning what should be a braze into a fusion weld.


That’s also why changing the pulse frequency or arc-length trim doesn’t fix the overheating issue.





So it’s not that my machine has corrupted firmware or missing menus — it’s behaving exactly as designed.


The user interface looks similar to the X-series, but under the hood it’s a locked-down waveform system with trim-level adjustments only.


Great for general fabrication, but not suitable for proper CuSi3 → Fe GMA-brazing or for anyone expecting full double-pulse control."

hope that makes sense, its late, I can expland on anything above tomorrow as I'm shattered...

I'm also happy to be proven totally wrong and an Idiot, I can't rule that out as a possibility

 

[superclarkey]

also, to be clear, the X5 package was essentially the same price, this was just presented as a compact version which made more sense for my use case.

But why would anyone by this over an X5 which has everything this lacks...

3ph machines are nearly always in workshops not at homes, the overlap makes no sense in the product line.

if I had been told the difference I would have gone the X5 route, and the fact I wanted it for the pulse to do CuSi3 to Fe is where my stems from...

 

[superclarkey]

wow I shouldn't reply when drinking lol...

 

[Munkul]

Have you actually tried a welder that DOES do what you expect it to do?
I'm a little bit confused if I'm honest. I don't think any welder make has a specific "cusi to fe" synergic line.

The cusi line on my machine, and must synergic machines tbh, works just fine for joining steel, full pen butt welds on thin material without overheating anything.

I get the feeling of being cheated if the pulse processor doesn't have the sophistication of the bigger machines, but how do you even know that? Again, you've had some pretty poor advice from sales people so far. Chinese whispers aren't helpful for this sort of thing.
And you should always judge a welder on usability and results, not the technology inside it.

I suspect a lot of your issues are down to bad feeding, due to the cooler being broke, and no one with enough skill/experience to show you how to get the best out of it.
Sorry mate, just my tuppence. If you were closer I'd love to have a play with it to see for myself if it's irredeemable or if it just needs some fixing.

You should absolutely keep pestering kemppi UK and shout loud and wide about the updates, wireless control, freezing issues etc. That sounds 100% their fault.

 

[superclarkey]

Let me be clear, the reason I have got upset is because I borrowed a machine that could do these things and realised that the limitations of mine wasn't in my mind, I then went down a rabbit of hole of learning about differences and thats when I came to learn that this machine (358) doesn't have sensing closed-loop and instead looks up a static table.

This is the issue, it went from a machine that was at a value price point that offered big hitting features to an over priced gimmick that is no better than the welders I had previously.

I don't believe there is any specific program for the Kemppi range for CuSi3 to FE, but doesn't mean that is 100% my fault, it was the main reason I wanted the machine, and pulse. SO you can appreciate why it annoys me.

I have now tried a Fronius and Kemppi X5 and both are worlds apart.

I'm upset as I'm not someone that falls into marketing ******** often, if at all, and there was no way of me knowing these subtle but very critical differences on face value as its clearly not highlighted.

The fact it has all the professional wifi and 4.0 ready stuff in it, (whatever that is) it feels to me like it was marketed towards professionals following WPS.

I understand that technically the high end machines are still CV and just use a close loop sensing feedback to mimic constant current, but I have had to learn this while trying to work out why my machine falls short.

Trying to explain this, here is what I read on the matter from ChatGPT (which I'm using to try and get to the bottom of my core issues here).. On the Kemppi the C01 is the CuSi3 welding program.

"
Why C01 (CuSi3 brazing) on the Master M 358 can’t just be fixed with trim settings

The C01 line on the Master M 358 is only a static open-loop synergy program — it’s not an adaptive waveform. The machine doesn’t measure or react to what’s actually happening in the arc. All it does is scale a pre-set lookup table of wire-feed vs voltage.

That means when you adjust the trim, you’re just shifting the entire voltage curve up or down — the pulse shape, peak/base current ratio, and timing never change. So if the arc conditions vary (different stick-out, part angle, or heat soak), the machine can’t compensate, and the process instantly falls outside the narrow window CuSi3 needs for proper wetting on steel.

The X-series Kemppi machines, or Fronius CMT/EWM ColdArc systems, have closed-loop adaptive control. They actually monitor arc voltage and current thousands of times per second and change the pulse waveform dynamically. That’s why those machines can keep the CuSi3 filler flowing evenly on steel — they’re regulating the heat input in real time.

On the Master M 358, C01 is a fixed waveform originally intended for CuSi3-to-Cu or generic brazing, not CuSi3-to-Fe. You can trim voltage or tweak inductance a little to soften the arc, but you can’t make it behave like the adaptive process because the control hardware simply isn’t there.

So the issue isn’t that the parameters are “off” — it’s that the machine has no feedback loop, so no amount of trimming can make it reproduce the correct CuSi3→Fe brazing waveform."
Also, with not having the correct gas, because far as I'm aware BOC doesn't offer it, and I have always used Argosheild Light and universal here, this is what ChatGPT had to say about that in relation to my machine.

"
Why being on the wrong gas mix (and trying to trim for it) causes problems on an open-loop machine like the M 358

Gas composition changes how the arc transfers metal — it alters ionisation potential, arc voltage, and cooling of the droplet. For example, switching from pure Argon to an Ar–CO₂ blend increases the arc voltage for the same current, changes droplet detachment energy, and shifts the transition point between short-arc, globular, and spray.

On a closed-loop adaptive machine (like the X5 or Fronius CMT), the control software measures this change in real time and automatically adjusts pulse frequency, current rise time, and base current to keep the arc stable and the heat input consistent.

Your M 358 doesn’t do that — it’s open-loop. The voltage and current are fixed from a lookup table that assumes the exact gas mix the program was written for (e.g. pure Ar for CuSi3). If you use a different gas (say 8 % CO₂ Argoshield Light) and try to trim voltage to compensate, all you’re really doing is shifting the entire curve up or down. You’re not correcting the pulse energy or the arc physics, so you end up outside the intended transfer mode — usually into globular or unstable short-arc — and the filler either overheats or fails to wet properly.

In short:

On a machine with no feedback loop, trimming can’t correct for gas-mix differences — it just changes voltage, not how the current behaves inside each pulse.
On an adaptive machine, the control firmware senses the change and reshapes the waveform automatically

"
I'm not saying ChatGPT is correct on every point, but I'm saying to you, I have been chasing issues in my setup and only when I borrowed other machines to see if I was the problem and having a simple out of box and work experience, it heavily points towards the Kemppi 358 not being suitable for my use case.

I will also argue if I was welding straight runs in manufacturing then I'm sure its likely not a problem, and problems get lesser as you are on 1.0 or 1.2mm wire, but if your welding tube (small 2inch tube) and you can't maintain "perfect" position as your trying to get as far around as possible, and or, out of position, or can't get the touch into a perfect CTWP you need a welder that can "sense" this and correct it, the Kemppi 358 has no such features and lacks what I thought it had, what it was meant to be, what it looked and tasted like, what I bought was a machine with programs that unless you stay strictly in the window of operation it will fall apart quickly and no work as expected.

I liken it to a vehicle that is mapped on a dyno and is tuned and still has closed loop for different situations, fuel, altitude etc, where as the 358 is like taking that map from the tuned vehicle and flashing it onto a car with no such closed loop.. as soon as you put bad fuel in, or go up a mountain it falls apart and the engine suffers. both might be same car, but one is truly taking advantage of the tools it has, the other doesn't care if conditions change it just provides the same instructions. (hope that makes sense).

So yes, it comes down to I feel like Kemppi wasn't straight up about what this machine is and what it can't do. Limitations of the programs are massive, when you understand how it works.

Feels like a money grab.

I'm likely be overly emotion about it, but I have walked past teh welder each day and opted to rent a machine or borrow one every time something like a cage comes up I need to do, I had a Kemppi Kempact 253A previously, which was a good welder from its price point, which was around £2k, this is the problem I have, this new welder is NO BETTER, it does the same things that welder did, you think PULSE and DOUBLE PULSE would be a game changer, but they are what feel like locked down in programs that again have no sense feedback and as such can start spitting like mad for no reason (specially if you CTWP or angle of touch becomes less than ideal).

I would love to be proven wrong, and I wish I had spoken up earlier, but Kemppi took my questions and would have known why I was running into this issues and just been clear and straight and said, well it can't do that because it doesn't have sense technology in it, that is how we made this machine cheaper... instead of years of thinking, why is this so crap.

ALSO, I had a chap in for a few days and he was welding T45 frame for me on my fixture table, I said nothing to him and said you can use my welder, he was messing and tweaking it all day, he said at the end of the day, I'm going to bring my welder in tomorrow, something is odd with that machine. he runs a Fronius, unsure what model but it looked basic and beaten up, I tried that on pulse and omg it was like getting out of a punto into a GT3 race car. Just worked, and I think that is the point, it should just work. He even tried his touch on my machine to see if it was a problem with that, and it was the same.

Again, he didn't say it was unusable, but he, like me, are judged on our welds from a cosmetic point of view just as importantly as being structurally sound, so if it looks like ass, customers think its ass.

I think I'm angry because I've always been made to feel like I'm imagining it, Kemppi saying it should do it, knowing full well it will never be able to do it as its missing important sensing closed loop set up within the machine, that is where I have lost all respect for Kemppi.

I'm not an expert and I'm not saying my finds are set in stone, I had a problem I couldn't seem to put my finger on, worked with them to try and fix it, parked the machine and when I started really looking into it, my symptoms match that of what I would expect of a machine pretending to be something its not.

Sorry for the rant.

 

[superclarkey]

Also if you look at the document that was available before I purchased the machine

https://www.rapidwelding.com/files/MSM358G.pdf

It shows software that is identical to X5 and higher models that could be purchased, so any reasonable person would think if this machine uses the same software add ons and uses for example the "WisePenetration+" which uses sense, closed loop to keep a constant heat input regardless of CTWP distance, you would assume it has the closed loop system to run that software.

However the software is different for the two machines, same price but one doesn't have the sensing capability, but to someone like me looking at the documentation before buying and seeing they run the same software upgrade programs you can see why I thought they had same tech inside.

Kemppi X5 WisePenetration+ Software (All X5) - Welding Supplies from Rapid Welding

Looking for...... X5500002, Kemppi X5 WisePenetration+ Software (All X5)

www.rapidwelding.com

Kemppi MSM WisePenetration+ Software (Master M 353, 355 & 358) - Welding Supplies from Rapid Welding

Looking for...... MSM50000, Kemppi MSM WisePenetration+ Software (Master M 353, 355 & 358)

www.rapidwelding.com

Looking at Kemppi own documents and the video showing the torch stepping away and adjusting, I fail to see how the 358 can do this without the sensing tech inside it. (its only sense is for display purposes in the from of a shunt at output stage).

WisePenetration MIG/MAG welding processes

WisePenetration is a welding process for ensured penetration in synergic MIG/MAG welding.

www.kemppi.com

So again, both listed to be able to do this, why wouldn't I believe they both have the tech inside.

 

[Brad93]

CuSi. Best machine you can buy for the job (non pulse speaking) is a ESAB CaddyMig or Esab U82 with the Qset feature.

I'm sure Fronius probably make something that does the same but I despise modern Fronius MIGs. Too complicated and then keep doing stuff you've asked it not to do in the background.

The only other thing I can compare Qset to is Millers with the RMD function although that is more advanced. qSet is very basic but very good at what it does.

 

[Munkul]

So you've compared against a good machine, nice. At least you know what it's supposed to be like.

Apart from chatgpt, what other knowledgeable source has told you the machine doesn't have adaptive arc sensing?
I say that, not in a disbelieving way, but because its something that's expected on any decent pulse machine these days, and for near £10k I'd absolutely expect it. It doesn't make sense for kemppi to put a cheap controller in and cause mayhem and disappointment.

Fronius and lorch and I'm sure others have had adaptive arc length for over a decade now, potentially multiple decades.