CycleOps Magnus vs Hammer smart trainer review

Don’t miss out on the latest CyclingTips updates.

Jump To Comments

The Magnus and Hammer are the newest additions to CycleOps’ range of indoor trainers. Both feature electromagnetic resistance units and broadcast speed and power measurements via ANT+ FE-C and Bluetooth 4.0, however the Magnus is a traditional rear-wheel-driven trainer while the Hammer is direct-drive.

In this review, Australian tech editor Matt Wikstrom takes a close look at the new smart trainers and explores the strengths and weaknesses of each.

CycleOps has been making indoor trainers for over two decades. Founded in 1995, the brand had some early difficulties with its fluid trainers before the Saris Cycling Group acquired it in 1999. The new owners overhauled the fluid trainer for 2001 and then released their first magnetic trainer in 2002.

After that, the company was quick to introduce electronics to its trainers, starting with the fully-wired eTrainer that was released in 2005, followed by the wireless PowerBeam Pro in 2008. Both models were among the earliest “smart trainers” where electronics controlled the resistance of the trainer to deliver a pre-programmed workout.

While those early trainers (Fluid2, Magneto, PowerBean Pro) are still part of CycleOps’ current catalogue, the company has moved on to add increasingly more sophisticated trainers to its line-up. Part of that growing sophistication takes advantage of technology from PowerTap, a brand that joined the Saris Cycling Group in 2008, so that users can monitor their power output and use it to structure their workouts.

The latest smart trainers from CycleOps: the Hammer (front) and the Magnus (rear).

CycleOps’ newest smart trainers, the Hammer and the Magnus, both make use of PowerTap technology. They also feature electromagnetic resistance units and share the same kind of wireless connectivity (ANT+ FE-C and Bluetooth 4.0) that is essential for a smart trainer. However they are quite distinct products that sell for very different prices.

An overview of the Hammer

The Hammer is CycleOps’ first direct-drive unit and sells for AUD$1,649/US$1,200/£1,200. It is designed to provide up to 2,000W of resistance and simulate slopes up to 20% while maintaining an accuracy of ±3% for all power measurements. The unit also measures speed, but not cadence.

One of the major benefits of any direct-drive trainer is that it saves wear on the rear tyre. For the Hammer, a precision-balanced 20lb/9kg flywheel takes the place of the rear wheel to provide a realistic pedalling action free of vibration and excessive noise.

A standard 11-speed Shimano/SRAM-compatible freehub body is provided for driving the unit along with axle fittings to suit quick-release (130/135mm) and thru-axles (142/148mm). As a result, the Hammer is compatible with a wide range of bikes however buyers must supply their own cassette and skewer/thru-axle.

The Hammer is a compact unit with support legs that fold into the base of the unit. In this state, the trainer stands 47cm tall, 22cm wide and 50cm long while unfolding the legs increases the width to 79cm. The Hammer is heavy, though, weighing over 21kg (the test unit weighed 21.9kg without a cassette or skewer) but there is a carry handle integrated into the unit that makes it easier to lift and move around.

A power cord (3.5m in length) is supplied with the unit along with a tray for the front wheel. The latter is designed to nestle in between the closed legs when the unit is packed away. Rider weight is limited to 300lb/136kg while CycleOps’ warranty covers the electronics for one year and the frame for a lifetime.

An overview of the Magnus

The Magnus is a largely traditional rear-wheel-driven trainer that is highly portable and costs AUD$760/US$600/£600. The unit provides a maximum resistance of 1,500W and can simulate gradients up to 15%, while the margin of error for power measurements is claimed to be ±5%. Like the Hammer, the Magnus measures speed and power, but not cadence, and broadcasts the data via ANT+ FE-C and Bluetooth 4.0 protocols.

The Magnus makes use of CycleOps’ classic frame that can be adjusted to suit 120mm, 130mm and 135mm hubs. A quick-release skewer is supplied with the unit for a sure fit with the frame. For those hoping to use a bike fitted with a rear thru-axle, there is an optional kit to suit many 12x142mm and 12x148mm axles.

The resistance unit can be adjusted to suit a range of wheel sizes from 26”-29” including 700C, however tyre width is limited to 2.0”. CycleOps’ provides its “clutch knob” for tensioning the resistance unit against the rear tyre, which clicks and slips like a torque wrench once the limit is reached. It’s a mechanism that ensures the rear tyre won’t slip on the drum of the unit and that power measurements will be consistent from one session to the next.

The Magnus is reasonably light (the test unit weighed 9.33kg), and thanks to the folding support legs and the hinged resistance unit, quite compact. At its smallest, the unit is 50cm tall, 49cm wide, and 36cm long; once set up for use, the unit measures 40cm tall, 78cm wide, and 68cm long.

The Magnus makes use of the same power cord as the Hammer that is 3.5m long. It also has the same rider weight limit (300lb/136kg) and warranty (one year for the electronics, a lifetime for the frame).

Getting started with each trainer

The Hammer and Magnus are supplied with almost everything that buyers will need to get started in the box, however some assembly is required. In the case of the Hammer, a cassette must be fitted, requiring a chain whip and lockring tool. This is a reasonably simple job but it can defeat the uninitiated.

Once the cassette is in place, the Hammer is ready for any road bike with a standard 130mm quick-release. For those using a 135mm quick-release or thru-axles, one or both of the axle end caps will need to be swapped out using a 5mm hex key and an adjustable wrench according to CycleOps’ instructions.

A little more work is required to assemble the Magnus: first, the resistance unit has to be fitted to the frame with the supplied nut and bolt; second, the clutch knob installed; third, the bolt-action tube fitted; and fourth, the hub-spacing set to suit 120/130/135mm hubs. For those using a thru-axle, an alternative bolt-action tube (supplied with the thru-axle kit) will need to be fitted to the frame.

Aside from the difference in drive mechanisms, there’s the weight, too. The Magnus (right) weighed in at 9.3kg versus 21.9kg for the Hammer (left).

Mounting a bike on the Hammer is much like fitting a rear wheel. The process is a little more cumbersome because the whole bike must be manoeuvred onto the axle of the unit, but once the dropouts are engaged, the quick-release skewer (or thru-axle) need only be tightened. Once this is done, the Hammer is ready for use.

The process is a little more involved for the Magnus, but for those that have used a rear-wheel-driven trainer before, entirely familiar. CycleOps’ quick-release skewer (or thru-axle) needs to be fitted to the rear wheel first, then the clutch knob unwound, the resistance unit lowered, and the left-hand shaft set for the hub width of the bike. As long as the bolt-action tube is open, the bike can then be positioned within the frame so that it can be closed onto the skewer/axle.

Closing the bolt-action shaft is just like closing a quick-release: the first-half of the action is easy, while the second-half requires some muscle. This can be tricky while trying to keep the bike aligned within the frame. Once the bike is secured, the resistance unit needs to be tensioned against the rear wheel by tightening the clutch knob until it clicks with the correct torque.

The effort required at this point is quite significant and users can expect to see the tyre squash up against the roller of the resistance unit. How much depends upon the size and pressure of the tyre. CycleOps makes no firm recommendations on the ideal tyre size or pressure for the Magnus, but for road tyres at least, I discovered that high pressures are better (see below).

Once a bike is mounted, the Hammer and Magnus are ready to use. Buyers opting for the Magnus will need some kind of tray for the front wheel, be it an old textbook, CycleOps’ levelling block, or even the company’s climbing riser block.

Both units can be used unpowered as simple resistance trainers for an easy workout or warming up before a race. Once powered though, owners will be able to tap into the full potential of each trainer provided they have a suitable device (smartphone, tablet, laptop, or desktop).

Getting connected

Very little happens when the Hammer and Magnus are powered up. There is an LED in each unit that comes on and changes colour according to its operating status. This sequence starts with a slow blinking green light, which indicates that the unit is ready. Once connected to a device via ANT+ FE-C, the LED changes to a solid white colour, or to blue, if Bluetooth 4.0 is being used. Any kind of red light indicates a problem with the unit.

Making that connection via ANT+ FE-C or Bluetooth 4.0 can be tricky, and in some cases, buyers may find that they’ll need additional hardware, depending on the device and the app they want to use. A reasonably current smartphone, tablet, or PC with Bluetooth 4.0 should be ready to connect with the Hammer and Magnus without the need for additional hardware.

For those that are new to smart trainers, an app is required to provide a virtual environment for riding the trainer and recording each session. There is a multitude to choose from, including Zwift, Sufferfest, Training Peaks, and CycleOps’ own VirtualTraining. All are free to use for an introductory period before a subscription is required.

CycleOps has versions of VirtualTraining to suit iOS, Android, and PC. I made use of the iOS version running on an iPad4 throughout the review period. The app was reasonably intuitive to use and I had no trouble connecting to either trainer via Bluetooth 4.0.

One of the reasons to consider using any brand’s proprietary software is that it is likely to provide critical support for calibrating the trainer. This is certainly true for VirtualTraining and fortunately there is no need for a current subscription to access this feature.

Similarly, VirtualTraining is necessary for updating the firmware for the Magnus and Hammer, at least for PC users. For those using an iOS device, they can download PowerTap’s free mobile app to take care of this chore, however CycleOps has yet to develop a suitable app for Android users.

Putting the Hammer and Magnus to use

Using the Hammer and Magnus trainers was really no more difficult than throwing a leg over the bike and pedalling. VirtualTraining offered me a choice of hundreds of routes from around the world and dozens of structured workouts. There was also the option of joining an online race, or, a free ride where I could manually adjust the resistance of the unit.

At moderate speeds (~35km/h), both trainers were reasonably quiet, but both emitted a steady whine. Compared to a background noise of 60dB, I measured 85-100dB for the Hammer and 90-100dB for the Magnus. At higher speeds, the Magnus could be louder, depending on the tyre and pressure I was using, and it could suffer from some pretty obvious vibration. By contrast, the Hammer was always calm and steady.

The Hammer provided a very smooth pedalling action that was quite realistic but there wasn’t any kind of “road feel” to add to the sensation. The flywheel was reasonably easy to get going and it maintained its momentum for a relatively long time. In some instances, it was too long, so I had to “coast” in between power-directed intervals, at which point VirtualTraining would auto-stop until I was able to apply measurable power to the pedals again.

This issue did not plague the Magnus at all. The rear wheel only ever required a few seconds to wind down. As a result, the pedalling action wasn’t quite as smooth, though that depended upon the tyre and air pressure I was using. Narrow tyres at high pressures were almost as smooth as the Hammer, while wider tyres at low pressures added some chop to my pedal stroke.

Calibrating each trainer was a simple matter using the VirtualTraining app. Once a route or workout was selected, the next page would display the trainer and sensors that were connected along with a button for calibrating the trainer. This would activate a brief routine involving getting the trainer up to speed (29-36km/hr) and holding it for ~30s before being instructed to cease pedalling to allow the unit to wind down.

The difference in the time taken to wind down added more time to this process for the Hammer, but the wait period was no longer than ~60s. As long as the trainer passed the calibration test, I was then free to advance to the next page to commence the session proper. By the end of the review period, it was clear that calibration was more important for the accuracy of the Magnus than the Hammer, but I routinely calibrated both trainers before every session.

My output power is very modest (sub-1000W) so I never came close to challenging the maximum resistance of either unit. Both trainers were very steady during my vigorous efforts, though I found it was possible to make the Hammer rock a little, while the Magnus was immovable. At face value, the Magnus might be better suited to bigger, more powerful riders, but I wasn’t heavy, or strong, enough to test this notion.

I tackled a few of the climbs in VirtualTraining’s catalogue and discovered that the Hammer did a fine job of simulating gradients up to 20%, steadily increasing the resistance each time the road steepened. By contrast, the Magnus only managed the same with slopes up to 10% then resorted to artificially slowing my speed for steeper gradients rather than adding to the resistance of the unit.

While the Hammer had more scope for modifying the resistance of the unit during the course of the climb, it still only approximated the sensations involved. Any spell that was conjured by the streaming video footage was easily broken because the angle of the bike didn’t change with the terrain and the effects of gravity were completely absent.

But then, that is the nature of indoor training, so it’s not really fair to criticise any trainer for failing to reproduce all of the sensations associated with riding a bike in the real world. Compared to earlier generations of indoor trainers, CycleOps’ smart trainers are a vast improvement, and when combined with an engaging app, should erase at least some of the sterility and boredom associated with indoor training.

In this regard, VirtualTraining provides a rich environment for engaging the user. There are a multitude of workouts and routes to explore that can be searched and ranked according to various criteria such as duration/distance, training score, gradients, and user ratings. In addition, some routes and workouts have an option for streaming video (including a selection from the Sufferfest), but the feature I enjoyed most was using Google’s global map to search for routes. I became a virtual world traveller zooming and scrolling from one country to the next in search of my next ride.

For those using a PC, there is also the possibility of using VirtualTraining’s free route editor to create and share rides. Any GPS file can be used to create a route, and if available, video footage can be added. There is no need to actually ride the route to create one to use for the trainer, since there is an integrated Google map for plotting new courses.

I didn’t explore the compatibility of the Magnus and Hammer with third-party apps other than Zwift. Such compatibility has become almost obligatory in the new age of virtual training, so prospective buyers will be pleased to know that both trainers are Zwift-ready. However, a USB dongle will probably be required to connect laptops and desktops to the Hammer and Magnus via ANT+.

Finally, a note about the portability of each trainer: at 9.3kg, the Magnus was reasonably easy to move around. The hinge for each leg was quite stiff though, so a bit of muscle was required to fold the unit up for storage or packing into the boot of a car. By contrast, the legs on the Hammer were much easier to fold away, and while the unit was considerably heavier, the carry-handle was well placed, so it wasn’t awkward to move.

Testing the accuracy of the Hammer

I used a set of Verve Cycling’s InfoCranks to validate the power measurements reported by the Hammer over a series of sessions comprising both power-driven workouts and real-world routes. All of the InfoCrank’s data was collected via ANT+ using a paired o-synce cycling computer while the Hammer’s data was collected via Bluetooth using VirtualTraining running on an iPad4.

This testing quickly identified an issue for the first unit supplied for review — a demo model — because it was over-reporting the amount of power I was generating (Figure 1A). The unit was under-reporting peaks in my output and smoothing out the power measurements. There were outward signs of these faults, and the unit passed the calibration process every time. Updating the firmware did not rectify any of these problems, either.

Apollo happily replaced the demo unit with a new one, and the accuracy was much improved. In some sessions, power measurements from the Hammer closely matched those from the InfoCranks, including smaller fluctuations (Figure 1B), albeit with what appeared to be a minor (~1s) lag. However, there were still instances when the Hammer failed to measure sudden peaks in my output and it could be insensitive to smaller fluctuations at a steady pace (Figure 1C).

Figure 1: failings in the accuracy of the Hammer smart trainer. The power measurements from the Hammer (blue line) were compared to those from a set of Infocranks (red line) fitted to the bike for a series of training sessions using the VirtualTraining app. (A) The first Hammer unit provided for review consistently reported power measurements that were consistently higher than those from the InfoCranks. Note that the Hammer also failed to accurately report the spikes in output. (B) A second Hammer unit proved to be more accurate, with power measurements closely matching those provided by the InfoCranks during a hilly route. (C) The Hammer still had trouble reporting sudden spikes in power output during a power-based workout along with smaller fluctuations. (D) During a steep hill climb, the power measurements reported by the Hammer suddenly increased in relation to the InfoCranks as cadence fell below 40rpm. At cadences above 40rpm, the power readings from each device were in good agreement.

Perhaps more disturbing was the way that the Hammer’s power measurements suddenly drifted upwards by 25-45W on climbs with steep slopes (Figure 1D). This drift was associated with low cadences (<40rpm) and disappeared as the slope eased and my cadence increased. This was only something that I discovered upon close analysis of the data well after I had returned the Hammer, so I didn’t get a chance to test how strong this association was, nor can I comment on how severe the drift might be.   Such issues are likely to be off-putting for any buyer, especially considering that they occurred with a new unit running the latest firmware that was consistently calibrated at the beginning of each session. Nevertheless, I expect all of these issues will be addressed as the Hammer’s firmware is revised, however CycleOps has not provided a timeframe for these improvements.

Testing the accuracy of the Magnus

I carried out the same kind of validation for the power measurements from the Magnus as I did for the Hammer with one important variation: I also wanted to test what kind of impact the rear tyre could have by testing different tyres (brands and widths) and tyre pressures. Each time I changed a tyre, wheel, or the tyre pressure, I re-tensioned the clutch knob and re-calibrated the Magnus before collecting any data.

I established very quickly that tyre pressure could have a profound impact on the accuracy of the power measurements by the Magnus. The severity of this effect varied, depending on the brand and width of the tyre, but in general, the accuracy of the Magnus improved with an increase in tyre pressure (Figure 2A). However, I found at least one tyre (Continental GP4000S, 25C) where the reverse applied, such that the greatest accuracy was seen at a low pressure (60psi) (Figure 2B).

In either instance, the Magnus under-reported my power output. I was using a power-based workout for these tests, so when the accuracy of the Magnus was off, I quickly felt it in the legs (more effort was required to hit the target power for each stage of the workout). This behaviour was quite distinct from the Hammer, which consistently over-reported my power.

Figure 2: the accuracy of the Magnus is influenced by the rear tyre. The power measurements from the Magnus (blue line) were compared to those from a set of Infocranks (red line) fitted to the bike for a series of training sessions using the VirtualTraining app. A variety of tyres were fitted to the rear wheel for these sessions and inflated to 40-120psi, as shown. (A) Initial testing demonstrated that the accuracy of the Magnus’ power measurements was strongly influenced by tyre pressure. At lower pressures (eg. 60psi, as shown), tended to report lower values for power output when compared to the InfoCranks. As tyre pressure was increased, the accuracy of the Magnus improved. (B) There were some tyres (eg. 25C Continental GP4000S, shown) where the accuracy of power measurements from the Magnus decreased as tyre pressure was increased. (C) An update for the firmware (v31.032) largely rectified this issue, however it was possible to find a tyre and pressure combination (28C Hutchison Sector, 40psi) where the Magnus did not match the InfoCranks. (D) A closer look at the power data from the Magnus after the firmware upgrade, showing reasonably good agreement with the InfoCranks over a wide range of measurements for a 25C tyre inflated to 100psi. (E) Using the Magnus in cold temperatures can result in a minor loss of accuracy if a “hot calibration” isn’t performed after a few minutes of use. Note how power measurements from the Magnus show a small increase after ~7min that is most obvious at low wattages.

This issue turned out to be something that CycleOps was well aware of, and indeed, was in the process of creating a new version of the firmware for the Magnus to address these inaccuracies. There was a wait of several weeks, but once installed, the new firmware improved the accuracy of the Magnus for a wider range of tyre pressures and widths.

It was still possible to find a tyre and pressure combination where the accuracy of the Magnus suffered (Figure 2C), but overall, the performance of the trainer was more robust. And compared to the Hammer, the Magnus was more sensitive to minor fluctuations in power and better able to capture sudden accelerations, albeit with a minor (~1s) lag (Figure 2D).

Just like many other rear-wheel-driven trainers, it was important to perform a hot calibration after a few minutes of using the Magnus. If I neglected to do this, the accuracy of the unit would start to drift as the tyre warmed up (Figure 2E). This effect was most obvious on cold days, so strictly speaking, it may not always be necessary to perform a second calibration once a session is underway. VirtualTraining can be configured to automatically prompt the user to perform a hot calibration, so I’d recommend making use of this app for a short warm-up before diving into any riding/training/racing session with the Magnus.

Summary and final thoughts

The market for smart trainers has grown a lot in the last couple of years as indoor training has found a new lease of life with engaging apps such as Zwift. The Hammer and Magnus fit neatly into this market, generally matching expectations while offering a few thoughtful touches. And when compared to the competition, CycleOps’ pricing for each unit is pretty good, too (though how good really depends on which part of the world prospective buyers are located).

The Magnus proved to be a solid performer offering buyers a stable platform for workouts along with a reasonable amount of resistance. Out of the box, this trainer is well suited to buyers that own a road bike with a standard quick-release rear wheel and a phone/tablet running iOS (other bikes and devices may require additional hardware). The Magnus is highly portable and its power measurements accurate but those looking for cadence will have to add a compatible sensor.

Importantly, the accuracy of the Magnus’ power measurements was subject to the width and pressure of the rear tyre, but without any method of validation, buyers will be unaware of any inaccuracies associated with their choice of tyre and pressure. At least CycleOps has developed a clutch knob for reproducibly tensioning the resistance unit and revised its firmware to minimise the impact of this issue. What is missing though, is a firm recommendation for the best tyre and/or pressure combinations, however my testing indicated that narrow tyres and high pressures were a better choice.

The Hammer was not subject to any of these complications (or tyre wear), and was arguably easier to use, too. It was also a little quieter, offered extra resistance and a more realistic pedal stroke, however the accuracy of the unit clearly needs some work. I expect the latter will be addressed with an update to the firmware at some point but prospective buyers are also likely to be disappointed by the fact that they’ll have to provide their own skewer and cadence sensor.

All of the issues identified for the Magnus and Hammer have to be considered somewhat surprising for a company that has almost 20 years of experience in manufacturing indoor trainers. I’m not prepared to dismiss either trainer, though, since the build quality appears very sound, and both trainers have a few thoughtful features. Instead, I see this is a case of hiccups that can afflict the first generation of any new product, especially those that are completely dependent on electronics.

CycleOps Hammer Gallery


CycleOps Magnus Gallery

What do each of the individual ratings criteria mean? And how did we arrive at the final score? Click here to find out.

Editors' Picks