Printed Circuit Board Assembly & PCB Design SMT Electronics Assembly Manufacturing Forum

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


Batch Ovens

Views: 4264

My company is looking into buying a batch reflow oven to run... - Nov 02, 2005 by jrr3434  

#37566

Batch Ovens | 2 November, 2005

My company is looking into buying a batch reflow oven to run quanties varing from 10 to 50 boards per week. Can any one tell me what are the limiations of these ovens? How well can they actually do lead free? What kind of through put can I expect? What are some good manufacturers? Any help would be appreciated.

Jay

reply »

#37568

Batch Ovens | 2 November, 2005

Hi,

We used a batch oven from Reddish Electronics out of the UK. It was convection, and ok, but had limitations.

Batch convection ovens start to heat up, so the metal of the oven is hotter and hotter and it effects the profile. So we ended up needing to watch the oven, wait for the reflow, and then count seconds to unload the oven.

Plus then how do you get rid of the heat. The problem in cool down meant we needed to lift the mesh the board sits on out of the oven gently to let the board actually cool down. If you leave the board in the oven, it just stays hot.

Then the airflow was not good. It was ok for the smaller cards we were doing, but when we did panels of 4, it started to discolor the edges but the middle still had not even reflowed.

So it was not a great solution, and we replaced it with a Quicky vapor phase oven.

Vapor phase was amazing, but the problem was tomb-stoning. We just never really got rid of it, and at the time we were doing about 500 boards a month, and we had to give up. We were using 0603, and larger parts would have been ok. The soldering is amazing, and if we were doing heavy products such as power supplies or something like that, I would go back to it.

So in the end we settled on a Soltec inline oven. If you wanted my recommendation after trying everything possible for batch ovens, it would be to never consider a batch oven, as I just don't think you can get an accurate profile.

New inline ovens are low cost now, and are worth every cent. Even a second hand one would be worth getting if it's clean. If you can get someone to loan you a profiler you will get a much better result.

The problem I see that batch ovens will never be able to solve is how to keep the profile repeating each time. The oven gets hot, and you cannot get rid of all the hot air after the run. If the profile is not correct it causes failures downline. When we put in our inline oven, we saw an immediate reduction in faults.

There might be some high tech batch oven out there, but with the cost of inline ovens, I would go inline, and never go back to batch.

I hope this helps.

Regards,

Grant

reply »

Slaine

#37583

Batch Ovens | 3 November, 2005

we had the same problem of getting rid of the heat from the oven. we set up an air line to blow into the oven after parts had been removed that helped a bit. I dont like using them because the profile is not the same every time.So where possible i put everything through a 5 zone btu instead.

reply »

#37591

Batch Ovens | 3 November, 2005

How well would a 3 zone oven work? I need to do lead free and have heard that 5 zones or more will work the best. The problem is I don't have the work load to justify a large oven. Any advice?

Jay

reply »


Rob

#37595

Batch Ovens | 3 November, 2005

Very badly in general.

You may get away with it if:

1) Your board if very thin & simple 2) Your heaters will actually allow the zones to reach the required temperatures 3) There are no fragile components on board 4) You've been very good & believe in Santa Claus.

Good luck.

reply »

JeffP

#37603

Batch Ovens | 3 November, 2005

Jay,

What are your board sizes? What is the maximum size that you envision reflowing? 10-50 boards a week . . . is that in a single run or broken down into a few boards a day/hour? Do you have a high mix or is it a single product? Are the boards double sided?

JP

reply »

#37609

Batch Ovens | 3 November, 2005

JP

We are a CM planing doing a high mix of boards ranging from 2"x4" to 8"x10" in a single run of 10 to 50. We don't plan on doing any double side boards.

Jay

reply »

#37610

Batch Ovens | 3 November, 2005

Jay,

As a CM you can forget about planning the types of boards you will be running. A CM needs to be flexible these days to stay in business. Forget the batch oven they are not flexible and have many limitations as pointed out earlier. Look for a 5 or more zone oven. A newer used oven would probably work for you and still not break the bank.

Jerry

reply »

#37624

Batch Ovens | 3 November, 2005

Thanks for all your help. I feel like I am headed in the right direction now. Thanks again.

reply »

#37633

Batch Ovens | 4 November, 2005

Hi,

Yes, you will save a lot of money in the end. From what I have experienced you cannot heat up a oven full of air, and cool it down fast enough to match a suitable reflow profile.

So passing the board through the zones using an inline oven is the only real way to go. I guess the confusing part is normally an inline machine is conveyered for speed, but with ovens it's really done to allow the temperatures to remain stable, and then pass the product through so you can get an accurate profile. It's also fast, but I think the main reason for inline is it's the only way to get an accurate profile.

It does not take much money in faulty products to pay for a used inline oven that's in good condition. Try for a good brand full convection, and be wary of the simple ovens, as you need temperature stability.

Regards,

Grant

reply »

JeffP

#37650

Batch Ovens | 4 November, 2005

Jay,

I work for a company that manufactures batch reflow ovens however, I will try to be as subjective as possible.

For the most part, what others have posted is true. Yes, there is one heating chamber and if the heater does not have enough power to ramp up rapidly then achieving an accurate profile with a high degree of repeatability would not be possible. It's just that not all batch ovens work the same way.

To answer your questions (as it relates to our oven): 1. Limitations? Throughput 2. Lead Free? Compatible . . . either ramp up to a high temperature and have the board sit a short period of time or stay longer at a lower temperature setting. Desired cooling rate is attainable. 3. Throughput? Typical cycle is 3-5 minutes long. The limitation is the loading tray size. For example, if the tray is 20"x20", you may reflow one board that is 20"x20" in 3-5 minutes or as many smaller ones that will fit in that array.

Grant mentioned that in his oven they had to count seconds from the moment the board enters the heating chamber to the time when the tray is pulled out for cooling. Thats because the heating chamber had to be heated prior to introducing the board. Programmable profile settings are limited to say the least. Cooling inside the chamber was non existent. That is certainly primitive and what others have described is true of batch ovens where the heaters cannot ramp up at the desired rate.

In our oven, you may program the ramp up rate up to 4 degC/second to a specific soak and reflow temperatures, the time you stay at those temperatures and the time at the cooling phase. The principle is that the board is stationary . . . the heater will ramp up to set temperatures. Cooling is achieved with a built in exhaust system that pulls out the hot air from the heating chamber (at a rate of 140 CFM) while ambient room air is cycled through the oven. Cooling may also be achieved by having the board/s exit the oven's chamber (no operator intervention) and by an external cooling module (fans directed at the tray). There is excellent thermal uniformity and a profile is consistent from batch to batch as you are starting reflow from the same temperature point. Considerably less floor space than a 5+ zone oven. High temperature rating, fast ramp up and cool down allow you the flexibility to running the same profile you would do in a 5+ zone oven but without the throughput. Profiling is very straightforward (for example, by modifying the rflow temperature setting you are only changing the profile in that section rather than from start to finish as in a conveyer oven if you are changing the belt speed). Also, keep in mind that you would not get far without a good profiler with a conveyer oven.

If you are interested, post your email and I will send some information on our product. However, main purpose of my reply was to point out that not all batch ovens operate under the same principle. Just like not all conveyer ovens will give you the flexibility and profile you want. BTW, don't even think of a 3 zone oven for lead free, or even eutectic on boards with a lot of thermal mass. An oven may be 3 zone, but if each "zone" is only 8" long you would essentially have the same temperature value at each of the top heaters. Also, look at the width of a conveyer as it will limit the width of your boards (if they are square).

JP

reply »

#37654

Batch Ovens | 4 November, 2005

JeffP

My email address is jrr3434@juno.com.

Jay

reply »

#37657

Batch Ovens | 4 November, 2005

Hi,

How do you keep the profile exactly the same as the walls of the oven heat up after dong a few boards?

That's the biggest problem with batch ovens, and as far as I can tell, there would be no way to keep a convection batch oven repeating the same profile exactly over many boards.

That's why we counted the seconds. We did not count from the time the board was introduced, because the oven did that. What we had to count was the seconds from the reflow point, because it would vary from board to board as the oven heated up.

We tried to keep it busy so it was hot, but it varied from board to board no matter what you do. I don't know how anyone could design a convection batch oven that's repeatable because the unit itself changes in temperature, and that causes problems.

That's why passing the board through the hot air like happens in an inline oven is the only way to get a repeatable profile. Without a repeatable profile that's correct, your asking for soldering problems and board failures.

Regards,

Grant

reply »

JeffP

#37658

Batch Ovens | 4 November, 2005

Hi Grant,

"How do you keep the profile exactly the same as the walls of the oven heat up after dong a few boards?"

You keep the profile the same by keeping the air temperature the same and more importantly, maintaining the same ramp up rate to the set temperature points. Yes, the metal inside the oven acts as a heat sink however, attaining a repeatable profile is a matter of air temperature repetition. If you are to attach a thermocouple to the walls of the heating chamber, after several reflow cycles the panels will be hotter than they were when you ran the first profile . . . so what? The air temperature in the oven is controlled by the internal sensor, as long as the heater has enough power to ramp up at a desired rate, whats the difference how hot the metal gets or that it gets slightly hotter from profile to profile. If we program an air temperature ramp up of 4 degrees C/second to a soak temperature of 180C, stay there for 60 seconds then ramp up at the same rate to 240C and stay there for 50 seconds whether its the first or the tenth profile, the air temperature curve will be the same regardless of temperature of the metal inside . . . heater will turn on and off intermittently to maintain your profile settings. I will say that the starting air temperature inside the oven during your second and subsequent reflow cycle most likely will be slightly higher than in the first cycle when the oven was cold. This however is insignificant. Typically, in the first cycle run the air temperature inside is that of room temperature (approx 25C) and around 60C at the start of profile 2,3,4,5,etc. depending on how long before the operator removes a soldered board and places a new batch. 60C is typical, it could be less if the operator runs a new batch after 1-2 minutes of the first cycle or a little more (something like 80C) if they run a second profile within a minute of completing the first. Regardless, from 25C to 180C at a rate of 4deg/C takes 39seconds and 30 seconds if you are starting at 60C. Nine seconds is insignificant (keep in mind that we are talking about air temperature not board temperature).

BTW, in a conveyer oven if you place one board followed by another say 2" behind, your profile on the second board may differ from the first board. This is a question of thermal load and how fast can the heaters in your oven recover(ramp back up to the set temperature). If they can't recover fast enough, then you will be exposing the second board to an air temperature that is now below that of the first board. In that case, there is no perfect repetition. Like in a batch oven, how fast the heaters can recover is critical to consistency in a profile. That's why we counted the seconds. We > did not count from the time the board was > introduced, because the oven did that. What we > had to count was the seconds from the reflow > point, because it would vary from board to board > as the oven heated up. > I am not certain exactly how your oven worked but I will take a shot at it. You program a specific air temperature, say 240C . . .oven ramps up to that temperature, you then open the cover/drawer place the board, by the time the door is closed, air temperature inside falls to say 200C. The oven then ramps up to your set point of 240. If the oven is relatively cold (your first cycle run) it may take 60 seconds, if this is your 10th profile, it may take 20 seconds since the metal inside is hot and does not absorb too much heat. Once at 240 it stays X amount of seconds that you program. So, I can see that in the first run, the total time that the board is inside = 60seconds + Xseconds at reflow and in your 10th cycle = 20seconds + X seconds. That 40 second difference is significant and I can see the inconsistency. Again, that is the question of the power of your heaters. Was the power consumption something like 2kw or 20kw? Cant see why you would need to count from 240C on regardless of the metal temperature. Once the set temperature is reached shouldn't the oven begin to count down your programmed seconds?

JP

reply »

#37667

Batch Ovens | 5 November, 2005

Hi,

I might not have been clear in my previous post on this, and the problem we saw is the hot metal of the oven walls caused IR transmission of heat into the board, and thats why the point in time from the start of the profile to the point the PCB went into reflow varied.

You ended up with an oven trying to use convection, but the oven itself was generating IR into the board based on how hot it become after cycles, and it's totally uncontrolled.

That's why I think batch ovens are never going to be any good. However I have not used your product.

As for the timing issue, the IR issue of the heat meant we could not predict when the board would get to reflow, so we had to wait and watch for the point when the solder turned shiny, and then count for the specified time the solder paste required above reflow.

Regards,

Grant

reply »

JeffP

#37669

Batch Ovens | 5 November, 2005

That was the case with your oven but it is not so with ours. First, the distance from walls of the inner chamber to the product differs from one oven to another. Second, the dissipation of heat from metal heats up the air in the oven which in turn heats the board. If the sensor that feedbacks/controls the oven is positioned properly inside, the controller would recognize the actual air temperature and adjust the heater accordingly to maintain your profile settings. When the oven is cold, the heater stays on longer/uses more power to maintain your programmed temperature ramp. When the metal inside gets hot, the heater uses less power because it mainly heats up the air considering that metal does not absorb as much heat as before. The net result is that the air temperature still ramps up at the same rate regardless of the metal temperature. It looks like your oven executed the profile settings without consideration for the actual air temperature (air heated up by the heat from the metal + heater). Otherwise, it would be consistent from cycle to cycle. Was the sensor reading actual air temperature or temperature somewhere between the heater and the inner chamber before it is forced into the reflow area?

JP

reply »

#37672

Batch Ovens | 5 November, 2005

Hi,

The oven we had used an air sensor, however we had heat from the metal oven walls IR conducting into the board. Once you run an oven, the metal retains a surprising amount of heat, and that radiates as IR energy out of the walls, and into the product. It, as you mention, also heats up the air.

I am interested in getting to know more about your oven, and it might be worth posting a url to the product specs. I am still very non trusting on batch convection ovens and them following a correct profile, however if you have worked it out it would be interesting to see.

Have you run PCB profiles on your oven repeatedly on a product with the same oven settings, and have plots of the temp/time graphs?

I would be interested to see the profile of a typical SMT board of the same product run 20 times, back to back, from a cold start, and then see the first profile and then the last.

Regards,

Grant

reply »

JeffP

#37725

Batch Ovens | 8 November, 2005

No question that metal retains a lot of heat (slow to heat up and slow to cool down). But again in a batch oven, it mainly boils down to the heater. I know, until you see proof you are skeptical.

I think by posting a url to product specs will open a can of worms followed by all kinds of accusations . . . "this is not a sales forum", etc. etc. As such, I better not:). Besides, it will not show anything relating to consistency of a profile.

Yes, we did run tests to measure profile repeatability. I am not sure whether we ran 20 profiles back to back but it certainly was a significant amount of profiles to make sure that the metal inside was at a high temperature. Can't locate the graphs from repeatability testing however, I will get some free time later this week/early next week and will perform the test. Can you post an email address and I will send the results to you? BTW, anyone who intends to run 20+ profiles back to back (less than 1 minute standby time from cycle to cycle) would probably be better off with a multizone (at least 4) oven in terms of throughput. Our batch oven is ideal for prototype - low volume runs. Greatest benefit is the flexibility . . . you may reflow some boards in the morning, shut the oven off, come back later in the day and reflow another batch/product without warming up the oven and essentially no set-up time. Those who are in R&D, prototype/testing labs, low volume OEM's, have a hard time justifying cost and allocating the space for a multizone oven.

On a side note, I located a graph showing temperature uniformity as measured with 5 thermocouples attached to a PCB. Let me know if you are interested.

JP

reply »

#37731

Batch Ovens | 8 November, 2005

Hi,

I would be interested in the data, and I after our experience with batch convection I am quite skeptical. I would be very interested in temperature plots.

What I would like to see before I ever considered a batch convection oven is how the oven can remain consistent over multiple reflows of various times between boards, just like you would use it when doing batch runs. So what is the profile on the first board, then on the following boards over the space of a few hours, some boards back to back, and some with a few minutes between them.

I would be interested in seeing how the profile held up and remained consistent. Anyone can do good airflow with a little brains, and so delta t is important, but not the main weakness I see on batch convection. That's why I would like to see the effects of IR into the PCB from the hot metal walls, as the walls change in temperature between boards over time.

I don't think anyone would mind you posting a URL, as then people reading the thread will know why you are. Other companies post all the time, which is annoying, but after you have put in so much work into the replies to this thread, it would be great to see more info on the products you mention here.

I think it's a good idea to keep the forums clear of people repeatedly marketing comments at the end of every thread, but we are not communists!

Please post when you have some temp plots!

Regards,

Grant

reply »

JeffP

#38122

Batch Ovens | 29 November, 2005

Grant,

Finally had some free time to run the profiles. Go to http://community.webshots.com/album/514175068ylUEdS

Temperature was measured with a thermocouple soldered (high temp alloy)to a 7" x 7" x 0.063" PCB. ProfileCycle1 was the first run in a cold oven. At the end of the first cycle, the board was removed and set aside to cool. The same profile was executed 9 more times consecutively to heat up metal panels inside the oven. Then the PCB was introduced in the next cycle (ProfileCycle10). I ran profiles continuously so that the metal inside would not get a chance to cool.

On the graph X axis represents time (10 seconds between dashed vertical lines and 30 seconds between solid lines) and the Y axis represents temperature in Celsius (5 degrees between 2 dashed horizontal lines and 20 degrees between solid lines).

Results: ProfileCycle1- a)25C to 120C in 50 seconds b)120C to 160C in 65 seconds c)160C to 219C in 60 seconds d)Peak temperature - 219C e)5 seconds at peak temperature f)Time above 183C - 90 seconds

ProfileCycle10- a)25C to 120C in 50 seconds b)120C to 160C in 50 seconds c)160C to 224C in 60 seconds d)Peak temperature - 224C e)5 seconds at peak temperature f)Time above 183C - 90 seconds

Since the metal inside was very hot during the last cycle, it did not act as much of a heat sink in that run than it did during the first cycle. Hence, the peak temperature is higher in the last cycle. I would think that your reasoning in that the metal radiates heat which in turn heats the board is incorrect . . . for this to occur, the temperature of the metal must be above the air temperature. In the very early stages of a profile that may be true however, for the metal to be at or above air temperature during soak and reflow stages of a profile the oven must be used back to back for a very long period of time without any stand by time mind you. This may be unacceptable for a CEM where every minute counts but the oven is not intended for someone with a streamlined process.

>>Anyone can do good airflow with a little brains, and so >>delta t is important, but not the main weakness I see on >>batch convection. Actually it is a main weakness with batch and small conveyer ovens on the market. There must be a lot of people with little brains out there or poor designs . . . I prefer the latter comment. When you have temperature variations of up to 20degreesC 5" apart, it is mighty tough achieving quality reflow across the board. It seems you have an expectation that a batch oven should run for hours back to back . . . in that case, I can see why delta T is not that important. A batch oven is intended for a user who does not have the volume and floorspace to justify a large multizone oven but still want the quality. BTW, I did post a graph showing temperature uniformity as measured at 5 points across the board.

I am not posting this to convince you to buy a batch. Just would like some feedback from industry experts out there.

JP

reply »

JeffP

#38437

Batch Ovens | 13 December, 2005

Grant,

Are you out there?

reply »

#38464

Batch Ovens | 14 December, 2005

Hi,

Cool, I did not notice your last post on this thread, and I will check out your results, as I am not convinced about batch convection, but if you've cracked it, and have a stable product, it's very good to know.

Regards,

Grant

reply »

Ryan

#43439

Batch Ovens | 21 August, 2006

JeffP (or anyone who received a link to his website),

Can you please send a link to the company's website to me? My email is:

reash@khcontrols.com

We are looking into batch ovens for prototype work.

Thanks! Ryan

reply »

table top X-ray inspection

ii-feed SMD Intelligent Feeder